Project Title: Predicting the Entrance Exam Outcome to IBA (Undergraduate Program)¶
Wali Ullah (ERP-09745)
Background of the study
Given the importance of higher education at individual, economic and societal level, it is imperative to investigate the factors that influence academic performance. The income disparity across countries and within a country necessitates inquiry into impact of socioeconomic differences (posh areas vs lower-class areas, peaceful areas vs areas with social unrest etc.) on academic performance. The gender of students is an important factor in determining academic performance. There is a general perception among masses that females are better in qualitative skills whereas males are better in quantitative skills however literature has been unable to establish the superiority of either gender with most research efforts revealing mixed results. Another significant factor that impacts academic performance in higher education institutes is prior qualification whether candidate is coming from private institutes or public, international curriculum or national, arts background or science etc.
This research project attempts to determine the relationship between the identified intellectual and non- intellectual factors and the extent of their impact on academic performance. Therefore, this research project aims to answer the following research questions:
Does socioeconomic background have a role in academic performance of students?
Does academic curriculum have any impact on quantitative and qualitative skills of students?
Does gender of students affect the academic performance?
Does female show better performance in qualitative skills whereas males in quantitative skills?
- Are female candidates more interested in BBA type programs whereas male in quantitatively rigorous programs such as BSCS, BSEM and BSACF?
In continuation of the fact that males tend to have a tilt towards quantitative side and females are dominant with qualitative or soft skills. The degree program of BBA & SSLA is dominated by female applicants. Whereas, the degrees such as BSACF & BSCS, namely the “good” degree as they branch out from science and technology, tend to be swept away by men.
Socioeconomic Background
The study uses location as a proxy for socioeconomic status of applicants. Since, most of the applicant pool for IBA is from Karachi we will divide Karachi into three groups based on the income status of the residents in those areas. For instance, all affluent areas of Karachi will be clustered in Karachi-1, middle income areas will be clustered in Karachi-2, and all the rest will be clustered in Karachi-3. Other cities are not being classified similarly.
• Karachi-1 includes DHA, Clifton, PECHS, Bahadurabad,
• Karachi-2- includes area Gulshan Iqbal, North Nazimabad, Saddar
• Karachi-3- includes all other areas of Karachi
Academic Curriculum
regarding the academic curriculum, we consider four different types of Boards (curriculum) from which the students come to IBA. At the SSC (qualification level 1) level these are: • O-Level, Matriculation, Agha Khan Board, and other international Boards And the HSC (Qualification level 2) level, these are: • A-Level, Intermediate, Aga Khan Board and Other international boards
To compute the performance score, each applicant’s score in mathematics and verbal section is divided by the total attainable marks of each section to obtain attained score. Research Question 1: Does socioeconomic background have a role in academic performance of students? (ANOVA-of score with city)
Socioeconomic status (SES) is believed to have a strong correlation with measures of academic performance; however, weak, and moderate correlations are frequently reported in the literature. This project uses location as a proxy for socioeconomic status of applicants. Since, most of the applicant pool for IBA is from Karachi we divided Karachi into three groups based on the income status of the residents in those areas. For instance, all affluent areas of Karachi are clustered in Karachi-1, middle income areas are clustered in Karachi-1, and all the rest are clustered in Karachi-3.
Research Question 2: Does academic curriculum have any impact on quantitative and qualitative skills of students? (ANOVA-of across board)
Students with O/A Level background are more equipped with quantitative and qualitative skills hence tend to perform better academically. Nevertheless, in some areas vocational qualification comes at par to A Level. Initially it was known that sound performance in mathematics has a direct connection to the attainment of a degree/graduation from college. However, for the science-based degree, prior knowledge in mathematics and desirable graduation result did not have a direct relationship.
Research Question 3: Does gender of students affect the academic performance? (ANOVA-gender)
Performance of both genders tend to vary specifically based on the following four categories: verbal skills, visual-spatial ability, mathematical ability, and aggression. Research clearly indicates that men outperformed women in subjects such as economics. Yet, at some instances, gender alone is a weak predictor of academic performance.
Research Question 4: Does female show better performance in qualitative skills whereas males in quantitative skills? (English math comparison-across gender graph) To compute the performance score, each applicant’s score in mathematics and verbal section is divided by the total attainable marks of each section to obtain attained score.
Females tend to be star performers in the components of qualitative elements. The left hemisphere dominance of the brain allows their verbal ability / speaking potential or qualitative element to be strong. On the contrary, male applicants are better off with hard skills / technical subjects that involve mathematics.
Research Question 5: Are female candidates more interested in BBA type programs whereas male in quantitatively rigorous programs such as BSCS, BSEM and BSACF? (Percentage comparison of male and females in BBA, ACF with CS and EM)
In continuation of the fact that males tend to have a tilt towards quantitative side and females are dominant with qualitative or soft skills. The degree program of BBA & SSLA is dominated by female applicants. Whereas, the degrees such as BSACF & BSCS, namely the “good” degree as they branch out from science and technology, tend to be swept away by men.
In [1]:
import TemplateML as template
from sklearn.model_selection import train_test_split
from sklearn.tree import export_graphviz
import pydot
from IPython.display import Image
In [2]:
FILE_NAME = "FinalData_testing.csv"
LABEL_COL = "test_successful"
df = template.load_data(FILE_NAME)
display(df.head())
print(df.shape)
print(df.dtypes)
| term_name | gender | date_of_birth | place_of_birth | postal_address | city1 | Province | city | countryname | seat_no | ... | cert_degree1 | discipline_Mat | medium | cert_degree2 | studied_maths | discipline2 | Eng_Score | Math_Score | Year | program | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | Fall 2014 | Male | NaN | Gilgit | Near CSD shop, Defense Colony, Jutial, Gilgit | Gilgit | Gilgit | Others | Pakistan | 801 | ... | Aga Khan Board | Science | English | Aga Khan Board | Yes | Science | 70.63 | 26.0 | 2014 | BSCS |
| 1 | Fall 2014 | Male | NaN | Mardan | Katlang road sultan muhmmad kali shanker mardan | Mardan | KPK | Others | Pakistan | 806 | ... | Matriculation | Science | English | Intermediate | Yes | Science | 45.63 | 52.5 | 2014 | BSEM |
| 2 | Fall 2014 | Male | NaN | Danyore,Gilgit | Tehsil Danyore,District Gilgit,Gilgit Baltistan | GilGit | Gilgit | Others | Pakistan | 807 | ... | Aga Khan Board | Science | English | Aga Khan Board | Yes | Science | 80.00 | 58.0 | 2014 | BSEM |
| 3 | Fall 2014 | Male | NaN | Gahkuch Paeen | Village and P/o Gahkuch Paeen Tehsil Punial, | Gahkuch | Gilgit | Others | Pakistan | 808 | ... | Aga Khan Board | Science | English | Aga Khan Board | Yes | Science | 69.38 | 51.0 | 2014 | BSCS |
| 4 | Fall 2014 | Male | NaN | D.I.KHAN | VILLAGE DURRIKHAIL P/O ATHOG TEHSIL PAHARPUR D... | D.I.KHAN | KPK | Others | Pakistan | 810 | ... | Matriculation | Science | English | Intermediate | Yes | Science | 60.00 | 23.5 | 2014 | BSEM |
5 rows × 23 columns
(38931, 23) term_name object gender object date_of_birth object place_of_birth object postal_address object city1 object Province object city object countryname object seat_no object test_center object test_successful object interview_successful float64 cert_degree1 object discipline_Mat object medium object cert_degree2 object studied_maths object discipline2 object Eng_Score float64 Math_Score float64 Year int64 program object dtype: object
In [3]:
df.columns
Out[3]:
Index(['term_name', 'gender', 'date_of_birth', 'place_of_birth',
'postal_address', 'city1', 'Province', 'city', 'countryname', 'seat_no',
'test_center', 'test_successful', 'interview_successful',
'cert_degree1', 'discipline_Mat', 'medium', 'cert_degree2',
'studied_maths', 'discipline2', 'Eng_Score', 'Math_Score', 'Year',
'program'],
dtype='object')
Data Set:
For any given research, the dataset plays the backbone role as it provides a collection of information based on specified characteristics. In this case, the data consisted of:
- Entry Test Scores of English (Verbal) Component
- Entry Test Scores of Mathematics (Quantitative) Component
- Success Ratio in the Entry Test (Yes/NO)
- Number of Years over which the Test was conducted (2014 to 2020)
- Specific attributes of the applicants, such as: a. Gender b. Age c. Prior Academic Background d. Area of Residence As a first off run at the cluster of information, the data set helped retrieve 29 attributes (columns/attributes) across 38,931 rows (observations).
In [4]:
display(df.head())
print(df.shape)
print(df.dtypes)
| term_name | gender | date_of_birth | place_of_birth | postal_address | city1 | Province | city | countryname | seat_no | ... | cert_degree1 | discipline_Mat | medium | cert_degree2 | studied_maths | discipline2 | Eng_Score | Math_Score | Year | program | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | Fall 2014 | Male | NaN | Gilgit | Near CSD shop, Defense Colony, Jutial, Gilgit | Gilgit | Gilgit | Others | Pakistan | 801 | ... | Aga Khan Board | Science | English | Aga Khan Board | Yes | Science | 70.63 | 26.0 | 2014 | BSCS |
| 1 | Fall 2014 | Male | NaN | Mardan | Katlang road sultan muhmmad kali shanker mardan | Mardan | KPK | Others | Pakistan | 806 | ... | Matriculation | Science | English | Intermediate | Yes | Science | 45.63 | 52.5 | 2014 | BSEM |
| 2 | Fall 2014 | Male | NaN | Danyore,Gilgit | Tehsil Danyore,District Gilgit,Gilgit Baltistan | GilGit | Gilgit | Others | Pakistan | 807 | ... | Aga Khan Board | Science | English | Aga Khan Board | Yes | Science | 80.00 | 58.0 | 2014 | BSEM |
| 3 | Fall 2014 | Male | NaN | Gahkuch Paeen | Village and P/o Gahkuch Paeen Tehsil Punial, | Gahkuch | Gilgit | Others | Pakistan | 808 | ... | Aga Khan Board | Science | English | Aga Khan Board | Yes | Science | 69.38 | 51.0 | 2014 | BSCS |
| 4 | Fall 2014 | Male | NaN | D.I.KHAN | VILLAGE DURRIKHAIL P/O ATHOG TEHSIL PAHARPUR D... | D.I.KHAN | KPK | Others | Pakistan | 810 | ... | Matriculation | Science | English | Intermediate | Yes | Science | 60.00 | 23.5 | 2014 | BSEM |
5 rows × 23 columns
(38931, 23) term_name object gender object date_of_birth object place_of_birth object postal_address object city1 object Province object city object countryname object seat_no object test_center object test_successful object interview_successful float64 cert_degree1 object discipline_Mat object medium object cert_degree2 object studied_maths object discipline2 object Eng_Score float64 Math_Score float64 Year int64 program object dtype: object
In [5]:
df.isnull().sum()
Out[5]:
term_name 0 gender 0 date_of_birth 3450 place_of_birth 1 postal_address 549 city1 0 Province 0 city 0 countryname 0 seat_no 385 test_center 11937 test_successful 0 interview_successful 1637 cert_degree1 0 discipline_Mat 0 medium 0 cert_degree2 0 studied_maths 0 discipline2 0 Eng_Score 0 Math_Score 0 Year 0 program 0 dtype: int64
Data Wrangling¶
Effective data management in today’s time has become such an important element of our profession that it allows us to oversee the execution happening according to the research’s true essence. Similarly, for the purpose of this investigation, candidate’s address was used to root down to their area, city and province details in SQL. In order to streamline the findings, rows with missing English and Mathematics scores were removed. Consequently, the final data that was closely considered and evaluated consisted of:
- 23 Columns (Consisting of the candidate’s details/specified attributes)
- 38,931 Rows (highlighting the number of observations made)
The dataset was composed such that it has two specific numeric columns populated by the scores in English and Mathematics. Moreover, one data column was specially dedicated the year in which the test was conducted along with the remaining 20 strings under scrutiny.
In [6]:
df = template.cleaningup(df, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['term_name','date_of_birth', 'place_of_birth',
'postal_address', 'city1', 'countryname','seat_no',
'medium','test_center','interview_successful',
'discipline_Mat'], to_date=['Year'])
df is all cleaned up..
In [7]:
df.columns
Out[7]:
Index(['gender', 'Province', 'city', 'test_successful', 'cert_degree1',
'cert_degree2', 'studied_maths', 'discipline2', 'Eng_Score',
'Math_Score', 'Year', 'program'],
dtype='object')
In [8]:
template.correlation_anlysis(df)
In [9]:
template.basicanalysis(df)
template.stringcolanalysis(df)
template.numcolanalysis(df)
Shape is:
(38931, 12)
Columns are:
Index(['gender', 'Province', 'city', 'test_successful', 'cert_degree1',
'cert_degree2', 'studied_maths', 'discipline2', 'Eng_Score',
'Math_Score', 'Year', 'program'],
dtype='object')
Types are:
gender object
Province object
city object
test_successful object
cert_degree1 object
cert_degree2 object
studied_maths object
discipline2 object
Eng_Score float64
Math_Score float64
Year datetime64[ns]
program object
dtype: object
Statistical Analysis of Numerical Columns:
Eng_Score Math_Score
count 38931.000000 38931.000000
mean 44.496195 35.989989
std 17.656205 17.088383
min 0.560000 0.500000
25% 32.500000 23.500000
50% 44.440000 36.000000
75% 56.670000 48.000000
max 97.780000 97.500000
The Descriptive Analysis of Data
The final data has a total of 38,931 rows (observations) and 12 columns (attributes). To further bifurcate the results the findings and analysis is mentioned herewith:
- The descriptive analysis of two numeric columns show that average English score is 44.49 with Standard Deviation of 17.65. Maximum and Minimum English score is 97.79% and 0.5% respectively. The average Mathematics score is about 36 with Standard Deviation of 17, maximum score of 97.5% and minimum of 0.5%. Moreover, it can be concluded that the average English score is higher than average Mathematics score, while both have similar volatility.
- The English and Mathematics scores boxplot show that there are some outliers. However, the two numeric columns English score and Mathematics score seems to be Normally distributed.
- The counts and percentage analysis of categorical variables show that:
a. Gender: 64% of the subject are male and 36% are females
b. Province: 88% of the applicants are from Sindh, while 12% from other provinces (which includes Punjab, AJK, KPK, Baluchistan and Gilgit etc)
c. City: Since Sindh populated the most number of applicants, it was evident because 50% candidates are from Karachi-3, 15% and 12% from Karachi 1 and 2 respectively. Hyderabad has 4%, Lahore 2% and Islamabad 2%, while other cities have less than 1%.
d. Test Success Ratio: Only 16% had the success ratio, whereas the remaining 84% of the candidates were unsuccessful as they failed the test.
e. Higher Education: 50% of applicants are coming from A/O level background, 3% from Aga Khan Board, about 47% from traditional Intermediate and secondary Boards and about 1% from other International Boards.
f. Studied Mathematics/ Science Background: 63% of the applicant have studied Mathematics in Intermediate and A level and about 54% of applicants are coming from Science background.
g. Program Selection: The applicant for BBA are 50%, BSACF are 25%, BSCS are 12%, SSLA are 6%, BSEM are 5% and BSECO are about 1% in the entire sample.
Normality Test of The Two Numeric Columns¶
In [10]:
template.Normality_test(df)
Normality Test for all numric columns ['Eng_Score', 'Math_Score'] Normaility Test for Column: Eng_Score Statistics=0.997, p_value=0.000 Sample does not look Gaussian (reject H0) Normaility Test for Column: Math_Score Statistics=0.991, p_value=0.000 Sample does not look Gaussian (reject H0)
The test-statistics for Normality show that both English and Mathematics scores are not Normally Distributed.
Mean Equality T-Test and Chi-Square Test of Independence¶
In [11]:
template.t_test(df)
template.chisquare_test(df)
t-test for equality of mean between all numric columns ['Eng_Score', 'Math_Score'] (Eng_Score,Math_Score) => t-value=68.30516117668779, p-value=0.0 Chisquare-test for Independence between all numric columns (Eng_Score,Math_Score) => chisqr-value=262956.6947531656, p-value=0.0 Dependent (reject H0)
The t-test suggests that mean score of English and Mathematics are not equal. However, the two scores are dependent of each other, subsequently, indicating that even though the English and Mathematics scores are not equal yet a candidate that is performing well in the English section also manages to score well in the quantitative section.
Percentage Counts Analysis of Categorical Columns¶
Research Question 5: Does female candidate are interested in BBA, SSLA and BSECO and type programs where as male in adcamically rigorous programs such as BSCS, BSEM and BSACF?
This hypothesis can be duly tested through the percentage count of male and female candidates in these two types of programs.
In [12]:
template.Count_Per(df, label_col=['gender'], cat_list=['program'])
==================***==================***==================
Analysis of Column gender
==================***==================***==================
=======================
Analsis of gender by program
=======================
%
program BBA BSACF BSCS BSECO BSEM SSLA
gender
Female 53.276925 19.862326 8.697835 1.183135 4.424208 12.555571
Male 48.949370 28.156894 14.472684 1.224735 4.446668 2.749650
<Figure size 360x1080 with 0 Axes>
In [13]:
template.Count_Per1(df, label_col=['gender'], cat_list=['program'])
Analsis of gender by program
In lieu of the research question mentioned above, the bar plot as well as the table show that females candidates are more inclined toward the BBA and SSLA program, whereas BSCS, and BSACF are male dominant programs. The percentage count of male and female candidates is almost same in both Economics programs (BSECO and BSEM). This result validates our hypothesis female applicants are more inclined towards BBA and SSLA type program, while male applicants towards the academically rigger type programs such as BSCS, and BSACF.
All comparisons in terms of percentage counts are given in Appendix-F with relevant explanations.
Average English and Mathematics Scores Comparisons¶
Research Question 4: Does female show better performance in English (qualitative skills) whereas males in Mathematics (quantitative skills)?
This hypothesis can be verified by comparing the average English and Mathematics scores of both male and female candidates for the two types of programs. (English math comparison-across gender graph)
In [14]:
template.Avg_by_cat(df,cat_list=['gender'])
=====================******=====================
Analysis of Average English and Mathematics Score by gender
=====================******=====================
Eng_Score Math_Score
gender
Female 46.492503 33.792770
Male 43.381906 37.216421
In [15]:
template.Avg_by_cat1(df,cat_list=['gender'])
Analysis of Average English and Mathematics Score by gender
The bar plot shows that average English score of female candidates are significantly higher than the male candidates, whereas of average Mathematics score of male is four times greater than the female candidates. Consequently, this verifies the Scientific research that women tend to display far superior potential when it comes to learning verbal abilities. They tend to rely more on this component as compared to any other strategy for easier recall. As opposed to gender differences in verbal ability favoring females, gender differences in mathematics performance favor males. Anastasi's (1958) text on differential psychology, states that differences in numerical aptitude favor boys.
Average English and Mathematics scores comparison across all string variables is given in Appendix-E.
ANOVA Analysis¶
ANOVA analysis is carried out to see whether the socio-economic status (denoted by the proxy “city”), curriculum (denoted by the proxy “Higher Education Board such as A-level, Intermediate etc.”) and gender has impact on English and Mathematics score of the candidates.
Research Question 1: Does socioeconomic background have a role in academic performance of students? (ANOVA of English and Mathematics scores with city)
This hypothesis can be empirically evaluated through ANOVA test of English and Mathematics scores using the string column city, which indicates the socio-economics classes in our sample.
In [16]:
template.ANOVA_analysis1(df,cat=['city'])
============+++++============+++++============ Analysis of Columns Eng_Score by city ============+++++============+++++============ Anova => - city
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 2.877417e+06 | 1.0 | 9450.781736 | 0.000000e+00 |
| C(Q("city")) | 2.857754e+05 | 8.0 | 117.327487 | 6.429085e-195 |
| Residual | 1.185032e+07 | 38922.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by city ============+++++============+++++============ Anova => - city
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.986099e+06 | 1.0 | 6808.878721 | 0.000000e+00 |
| C(Q("city")) | 1.480408e+04 | 8.0 | 6.344046 | 2.966714e-08 |
| Residual | 1.135325e+07 | 38922.0 | NaN | NaN |
The F-Statistics tend to be highly significant across socio-economics classes (city) for both English and Mathematics scores, which suggest that the income level and socio-economic class is an important determinant of the academic performance.
Research Question 2: Does academic curriculum have any impact on quantitative and qualitative skills of students? (ANOVA of English and Mathematics scores across Boards)
For further analysis of this hypothesis, it was empirically evaluated through ANOVA test of English and Mathematics scores using the string column cert_degree (which has the entries of Higher Education Board such as: A/O level, Aga Khan Board, Matric/Intermediate and other International Boards), ultimately highlighting the different types of curriculums in our sample. For this there are two variables cert_degree1 (it is curriculum type at 10 year of Education) and cert_degree2 (it is curriculum type at 12 years of education)
In [17]:
template.ANOVA_analysis1(df,cat=['cert_degree1','cert_degree2'])
============+++++============+++++============ Analysis of Columns Eng_Score by cert_degree1 ============+++++============+++++============ Anova => - cert_degree1
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 2.058587e+06 | 1.0 | 7410.358798 | 0.0 |
| C(Q("cert_degree1")) | 1.322233e+06 | 3.0 | 1586.560912 | 0.0 |
| Residual | 1.081387e+07 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by cert_degree1 ============+++++============+++++============ Anova => - cert_degree1
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.135856e+06 | 1.0 | 4553.895797 | 0.0 |
| C(Q("cert_degree1")) | 1.658690e+06 | 3.0 | 2216.683801 | 0.0 |
| Residual | 9.709370e+06 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by cert_degree2 ============+++++============+++++============ Anova => - cert_degree2
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 4.639299e+07 | 1.0 | 160595.265526 | 0.0 |
| C(Q("cert_degree2")) | 8.908121e+05 | 3.0 | 1027.886759 | 0.0 |
| Residual | 1.124529e+07 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by cert_degree2 ============+++++============+++++============ Anova => - cert_degree2
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.115302e+07 | 1.0 | 113883.053281 | 0.0 |
| C(Q("cert_degree2")) | 7.194733e+05 | 3.0 | 876.702924 | 0.0 |
| Residual | 1.064859e+07 | 38927.0 | NaN | NaN |
The F-Statistics are statistically significant across curriculums (for both certificate degree 1 and certificate-degree_2) for both English and Mathematics scores. Hence, enunciating the fact that that the curriculum which is taught at 9-10th classes and 11-12th classes is an important determinant of the academic performance in the entry test.
Research Question 3: Does gender of students affect the academic performance? (ANOVA of English and Mathematics scores across gender)
The hypothesis was empirically evaluated through ANOVA test of English and Mathematics scores using the string column gender.
In [18]:
template.ANOVA_analysis1(df,cat=['gender'])
============+++++============+++++============ Analysis of Columns Eng_Score by gender ============+++++============+++++============ Anova => - gender
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.014502e+07 | 1.0 | 97391.215552 | 0.000000e+00 |
| C(Q("gender")) | 8.660059e+04 | 1.0 | 279.785440 | 1.384166e-62 |
| Residual | 1.204950e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by gender ============+++++============+++++============ Anova => - gender
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.592565e+07 | 1.0 | 55044.080352 | 0.000000e+00 |
| C(Q("gender")) | 1.049090e+05 | 1.0 | 362.598437 | 1.779221e-80 |
| Residual | 1.126315e+07 | 38929.0 | NaN | NaN |
For which the F-statistics results proved to be highly statistically significant across gender (male and female) for both English and Mathematics scores, which suggest that applicant gender is an important determinant of the academic performance in the entry test.
Preparing Data for Machine Learning Algorithms¶
The one hot encoding is applied to string columns to create dummies, except the test_successful column which results in Rows=38931, and columns= 28). The data shows that 92% of applicants are from Sindh province. Therefore, the province variable is categorized in two group, Province_Sindh and Province_No_Sindh (all other provinces are groped in Province_No_Sindh).
In [19]:
df1=template.apply_label_encoding(df, cols=['test_successful'])
In [20]:
df1 = template.cleaningup(df1, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['Year', 'program'])
df is all cleaned up..
In [21]:
df1 = template.onehotencoding(df1)
In [22]:
df1.columns
Out[22]:
Index(['test_successful', 'Eng_Score', 'Math_Score', 'gender_Female',
'gender_Male', 'Province_AJK', 'Province_Balochistan',
'Province_Foreign', 'Province_Gilgit', 'Province_Islamabad',
'Province_KPK', 'Province_Punjab', 'Province_Sindh', 'city_Hyderabad',
'city_Islamabad', 'city_Karachi-1', 'city_Karachi-2', 'city_Karachi-3',
'city_Lahore', 'city_Others', 'city_Peshawar', 'city_Quetta',
'cert_degree1_Aga Khan Board', 'cert_degree1_Matriculation',
'cert_degree1_O-Level', 'cert_degree1_Other boards',
'cert_degree2_A-Level', 'cert_degree2_Aga Khan Board',
'cert_degree2_Intermediate', 'cert_degree2_Other boards',
'studied_maths_No', 'studied_maths_Yes', 'discipline2_Arts',
'discipline2_Science'],
dtype='object')
In [23]:
df1["Province_No_Sindh"]= df1['Province_AJK']+df1['Province_Balochistan']+df1['Province_Foreign']+df1['Province_Gilgit']+df1['Province_Islamabad']+df1['Province_KPK']+df1['Province_Punjab']
df1 = template.cleaningup(df1, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['Province_AJK', 'Province_Balochistan','Province_Foreign',
'Province_Gilgit', 'Province_Islamabad','Province_KPK', 'Province_Punjab'])
df is all cleaned up..
In [24]:
df1 = template.onehotencoding(df1)
In [25]:
df1.shape
Out[25]:
(38931, 28)
In [26]:
template.correlation_anlysis(df1)
Machine Learning Algorithms¶
Given the data set, there are three variables that are to be predicted
- Test_Successful, (Yes/No)
- Math_Score (Continuous Variable)
- Eng_Score (Continuous Variable) Followed by two regression type analysis on the following two elements that are closely considered:
- English Score
- Mathematics Score
Lastly, there is one classification-based scenario is considered, which is based on test_successful (1 for successful and 0 otherwise)
In the first step we try to predict the binary outcome variable test successful using the classification based algorithms. The binary variable test_sucessful is imbalance. The following command shows that 84% are test un-successful (zeros) and only 16% are ones. Therefore, it seems necessary to address the class imbalance. The application Random under sampling results in 12404 rows and 26 columns/attributes. In this stage we consider the Random Under Sampling to address the imbalancing issue. The results of ML algorithms with addressing the class imbalance issue are given here. The results without addressing the class imbalance issue are given in Appendix-A.
In [27]:
import seaborn as sns
import matplotlib.pyplot as plt
sns.countplot(x='test_successful', data=df1, palette='hls')
plt.show()
count_no_sub = len(df1[df1['test_successful']==0])
count_sub = len(df1[df1['test_successful']==1])
pct_of_no_sub = count_no_sub/(count_no_sub+count_sub)
print("percentage of test un-successful is", pct_of_no_sub*100)
pct_of_sub = count_sub/(count_no_sub+count_sub)
print("percentage of test successful is", pct_of_sub*100)
percentage of test un-successful is 84.0692507256428 percentage of test successful is 15.930749274357195
Classification based Prediction of Success with Addressing Class Imbalancing¶
In [28]:
S_all= template.cleaningup(df1, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['Eng_Score','Math_Score'])
df is all cleaned up..
In [29]:
S_all_bal1 = S_all.copy()
SAll_bal=template.Random_UnderSampling(S_all_bal1,'test_successful')
print('Orignal data shape:', S_all.shape)
print('Resampled data shape:', SAll_bal.shape)
Orignal data shape: (38931, 26) Resampled data shape: (12404, 26)
To predict the test outcome and select the best candidate ML algorithm, we have applied twelve (12) different machine learning classification algorithms and considered four different scenarios, which are:
a. Without focusing the cross validation (CV) and features selection
b. Only cross validation (CV: Stratified K-Fold) is considered
c. Only feature selection (random forest based algorithm is used) criterion is considered
d. Both cross validation (CV: Stratified K-Fold) and features selection (random forest based algorithm is used) are considered.
For features selection, we run the random forest based algorithm to extract the top candidate features.
In [30]:
#1. Classification based Algrithms Results without CV and RFFS and addressing class imbalancing
results_without_cv_reg_rffs= template.run_algorithms_clf(SAll_bal,'test_successful')
============ LogReg =========== Prediction Vector: [1 0 1 ... 0 0 1] Accuracy: 63.7646110439339 Precision of event Happening: 66.73306772908366 Recall of event Happening: 54.25101214574899 AUC: 0.6372261682728861 F-Score: 0.5984814649397053 Confusion Matrix: [[912 334] [565 670]] ============================== ============ KNN =========== Prediction Vector: [1 1 1 ... 1 1 0] Accuracy: 61.104393389762194 Precision of event Happening: 60.44891640866873 Recall of event Happening: 63.238866396761125 AUC: 0.6111381522085247 F-Score: 0.6181242580134546 Confusion Matrix: [[735 511] [454 781]] ============================== ============ GadientBoosting =========== Prediction Vector: [1 0 1 ... 0 0 1] Accuracy: 64.81257557436517 Precision of event Happening: 66.48451730418944 Recall of event Happening: 59.10931174089069 AUC: 0.6478740065375193 F-Score: 0.6258036862408917 Confusion Matrix: [[878 368] [505 730]] ============================== ============ AdaBoost =========== Prediction Vector: [1 0 1 ... 0 0 1] Accuracy: 63.804917372027404 Precision of event Happening: 66.76616915422886 Recall of event Happening: 54.33198380566802 AUC: 0.6376310265724813 F-Score: 0.599107142857143 Confusion Matrix: [[912 334] [564 671]] ============================== ============ SVM =========== Prediction Vector: [1 1 1 ... 1 0 1] Accuracy: 65.25594518339379 Precision of event Happening: 64.0542577241899 Recall of event Happening: 68.82591093117408 AUC: 0.6527170345916651 F-Score: 0.663544106167057 Confusion Matrix: [[769 477] [385 850]] ============================== ============ DecisionTree =========== Prediction Vector: [1 1 1 ... 1 1 1] Accuracy: 64.36920596533656 Precision of event Happening: 63.984063745019924 Recall of event Happening: 65.02024291497975 AUC: 0.6437207972394253 F-Score: 0.6449799196787148 Confusion Matrix: [[794 452] [432 803]] ============================== ============ DeepNeuralNetwork =========== Prediction Vector: [1 1 1 ... 1 0 1] Accuracy: 65.21563885530028 Precision of event Happening: 63.88059701492538 Recall of event Happening: 69.31174089068826 AUC: 0.6523371956251909 F-Score: 0.6648543689320388 Confusion Matrix: [[762 484] [379 856]] ============================== ============ RandomForest =========== Prediction Vector: [1 1 1 ... 1 1 1] Accuracy: 64.40951229343007 Precision of event Happening: 63.990461049284576 Recall of event Happening: 65.18218623481782 AUC: 0.6441292297294663 F-Score: 0.6458082631367831 Confusion Matrix: [[793 453] [430 805]] ============================== ============ NaiveBayes =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 60.25796049979847 Precision of event Happening: 61.226330027051404 Recall of event Happening: 54.97975708502024 AUC: 0.6023466184909119 F-Score: 0.5793515358361775 Confusion Matrix: [[816 430] [556 679]] ============================== ============ MultiLayerPerceptron =========== Prediction Vector: [1 1 1 ... 1 0 1] Accuracy: 65.05441354292624 Precision of event Happening: 63.96054628224582 Recall of event Happening: 68.2591093117409 AUC: 0.6506855947127975 F-Score: 0.6603995299647474 Confusion Matrix: [[771 475] [392 843]] ============================== ============ LightGbm =========== [LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines [LightGBM] [Info] Number of positive: 4967, number of negative: 4956 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000691 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 46 [LightGBM] [Info] Number of data points in the train set: 9923, number of used features: 23 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500554 -> initscore=0.002217 [LightGBM] [Info] Start training from score 0.002217 Prediction Vector: [1. 0. 1. ... 0. 0. 1.] Accuracy: 64.65135026199114 Precision of event Happening: 67.61811023622047 Recall of event Happening: 55.62753036437247 AUC: 0.64611517991175 F-Score: 0.6103953798311862 Confusion Matrix: [[917 329] [548 687]] ============================== ============ XgBoost =========== [11:18:09] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior. Prediction Vector: [1 1 1 ... 1 1 1] Accuracy: 65.01410721483273 Precision of event Happening: 63.62286562731997 Recall of event Happening: 69.39271255060729 AUC: 0.6503343492698903 F-Score: 0.6638264910921766 Confusion Matrix: [[756 490] [378 857]] ============================== ============ LightGbm =========== [LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines [LightGBM] [Info] Number of positive: 4967, number of negative: 4956 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000607 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 46 [LightGBM] [Info] Number of data points in the train set: 9923, number of used features: 23 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500554 -> initscore=0.002217 [LightGBM] [Info] Start training from score 0.002217 Prediction Vector: [1. 0. 1. ... 0. 0. 1.] Accuracy: 64.65135026199114 Precision of event Happening: 67.61811023622047 Recall of event Happening: 55.62753036437247 AUC: 0.64611517991175 F-Score: 0.6103953798311862 Confusion Matrix: [[917 329] [548 687]] ==============================
In [31]:
# 2. Classification based Algrithms Results with CV only and addressing class imbalancing
results_cv = template.run_algorithms_cv_clf(SAll_bal,'test_successful')
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000455 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000441 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000796 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000517 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000434 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000584 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000469 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000457 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000450 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000433 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[11:21:38] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:39] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:40] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:41] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:42] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:43] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:44] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:45] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:46] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:21:47] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000595 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000863 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000449 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000455 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000617 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000619 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000721 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000711 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000643 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000725 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 46
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 23
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
============ LogReg ===========
{'accuracy': 64.37242988224897, 'precision': 67.07863263547455, 'recall': 52.91283569684691, 'auc_val': 0.6437173393589946, 'f_score': 0.5839778783503413}
==============================
============ KNN ===========
{'accuracy': 59.54426061188947, 'precision': 59.104046963583166, 'recall': 60.137914913511, 'auc_val': 0.5954393278271259, 'f_score': 0.5931794848416684}
==============================
============ GadientBoosting ===========
{'accuracy': 62.58303007460165, 'precision': 63.00988916550974, 'recall': 57.07248974079268, 'auc_val': 0.6258194379512753, 'f_score': 0.58899699682759}
==============================
============ AdaBoost ===========
{'accuracy': 64.29982974188349, 'precision': 66.99204150951577, 'recall': 52.54186795491144, 'auc_val': 0.6429914030439978, 'f_score': 0.5806882448372633}
==============================
============ SVM ===========
{'accuracy': 61.87339164565517, 'precision': 61.769389593200025, 'recall': 56.99181860682562, 'auc_val': 0.6187238325281804, 'f_score': 0.5825271064557936}
==============================
============ DecisionTree ===========
{'accuracy': 61.09942554131684, 'precision': 60.60941935605579, 'recall': 56.3308399563659, 'auc_val': 0.6109864422627396, 'f_score': 0.5742343925007559}
==============================
============ DeepNeuralNetwork ===========
{'accuracy': 61.6719282056614, 'precision': 61.417439975742, 'recall': 57.55649057191836, 'auc_val': 0.6167095215832943, 'f_score': 0.5859762143650451}
==============================
============ RandomForest ===========
{'accuracy': 61.23646382989786, 'precision': 60.68578404660136, 'recall': 57.088826554464696, 'auc_val': 0.6123587086385123, 'f_score': 0.5791066046344249}
==============================
============ NaiveBayes ===========
{'accuracy': 58.23730212367758, 'precision': 56.64873229963943, 'recall': 53.782089242117294, 'auc_val': 0.5823722144304192, 'f_score': 0.5415082974746674}
==============================
============ MultiLayerPerceptron ===========
{'accuracy': 61.69608926204154, 'precision': 61.649833932727276, 'recall': 56.60472183263207, 'auc_val': 0.6169508077502467, 'f_score': 0.5796162442944286}
==============================
============ LightGbm ===========
{'accuracy': 61.0106963686933, 'precision': 61.55426734178381, 'recall': 53.00875279206275, 'auc_val': 0.610099864942081, 'f_score': 0.5577695719362594}
==============================
============ XgBoost ===========
{'accuracy': 61.74442437160459, 'precision': 61.453228834338084, 'recall': 56.86263051270064, 'auc_val': 0.6174366266687444, 'f_score': 0.5792053014885182}
==============================
In [32]:
## 3. Classification based Algrithms Results with RFFS only and addressing class imbalancing
res_rffs = template.MachineLearningwithRFFS_clf(SAll_bal,'test_successful', threshold=5,
algo_list=template.get_supported_algorithms_clf())
cert_degree2_Intermediate 16.765584 studied_maths_No 13.986408 cert_degree2_A-Level 13.199462 studied_maths_Yes 9.010187 cert_degree1_Other boards 5.807934 cert_degree1_Matriculation 4.979888 cert_degree1_O-Level 4.431025 discipline2_Science 3.817486 discipline2_Arts 3.125780 city_Others 2.932969 gender_Female 2.634058 city_Karachi-2 2.258575 gender_Male 2.216016 city_Karachi-3 2.064187 Province_Sindh 1.776447 city_Karachi-1 1.719548 Province_No_Sindh 1.577353 cert_degree1_Aga Khan Board 1.460297 city_Hyderabad 1.424049 cert_degree2_Aga Khan Board 1.176838 city_Islamabad 1.017953 city_Lahore 0.806581 cert_degree2_Other boards 0.781467 city_Peshawar 0.643353 city_Quetta 0.386554 dtype: float64 Selected Features =['cert_degree2_Intermediate', 'studied_maths_No', 'cert_degree2_A-Level', 'studied_maths_Yes', 'cert_degree1_Other boards'] (12404, 26) ============ LogReg =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.918178153970175 Precision of event Happening: 67.79661016949152 Recall of event Happening: 48.582995951417004 AUC: 0.6285490086495408 F-Score: 0.5660377358490566 Confusion Matrix: [[961 285] [635 600]] ============================== ============ KNN =========== Prediction Vector: [1 1 1 ... 1 1 1] Accuracy: 63.72430471584038 Precision of event Happening: 62.89453425712086 Recall of event Happening: 66.15384615384615 AUC: 0.6373502901592789 F-Score: 0.6448303078137333 Confusion Matrix: [[764 482] [418 817]] ============================== ============ GadientBoosting =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.87787182587666 Precision of event Happening: 67.88154897494306 Recall of event Happening: 48.259109311740886 AUC: 0.6281334277786081 F-Score: 0.5641268338854708 Confusion Matrix: [[964 282] [639 596]] ============================== ============ AdaBoost =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.918178153970175 Precision of event Happening: 67.79661016949152 Recall of event Happening: 48.582995951417004 AUC: 0.6285490086495408 F-Score: 0.5660377358490566 Confusion Matrix: [[961 285] [635 600]] ============================== ============ SVM =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.87787182587666 Precision of event Happening: 68.77990430622009 Recall of event Happening: 46.558704453441294 AUC: 0.628058369779245 F-Score: 0.5552873008208595 Confusion Matrix: [[985 261] [660 575]] ============================== ============ DecisionTree =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.87787182587666 Precision of event Happening: 68.77990430622009 Recall of event Happening: 46.558704453441294 AUC: 0.628058369779245 F-Score: 0.5552873008208595 Confusion Matrix: [[985 261] [660 575]] ============================== ============ DeepNeuralNetwork =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.87787182587666 Precision of event Happening: 67.88154897494306 Recall of event Happening: 48.259109311740886 AUC: 0.6281334277786081 F-Score: 0.5641268338854708 Confusion Matrix: [[964 282] [639 596]] ============================== ============ RandomForest =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.87787182587666 Precision of event Happening: 68.77990430622009 Recall of event Happening: 46.558704453441294 AUC: 0.628058369779245 F-Score: 0.5552873008208595 Confusion Matrix: [[985 261] [660 575]] ============================== ============ NaiveBayes =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.918178153970175 Precision of event Happening: 67.79661016949152 Recall of event Happening: 48.582995951417004 AUC: 0.6285490086495408 F-Score: 0.5660377358490566 Confusion Matrix: [[961 285] [635 600]] ============================== ============ MultiLayerPerceptron =========== Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.918178153970175 Precision of event Happening: 67.79661016949152 Recall of event Happening: 48.582995951417004 AUC: 0.6285490086495408 F-Score: 0.5660377358490566 Confusion Matrix: [[961 285] [635 600]] ============================== ============ LightGbm =========== [LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines [LightGBM] [Info] Number of positive: 4967, number of negative: 4956 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000115 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 10 [LightGBM] [Info] Number of data points in the train set: 9923, number of used features: 5 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500554 -> initscore=0.002217 [LightGBM] [Info] Start training from score 0.002217 [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf Prediction Vector: [1. 0. 1. ... 0. 0. 0.] Accuracy: 62.918178153970175 Precision of event Happening: 67.79661016949152 Recall of event Happening: 48.582995951417004 AUC: 0.6285490086495408 F-Score: 0.5660377358490566 Confusion Matrix: [[961 285] [635 600]] ============================== ============ XgBoost =========== [11:22:00] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior. Prediction Vector: [1 0 1 ... 0 0 0] Accuracy: 62.87787182587666 Precision of event Happening: 68.77990430622009 Recall of event Happening: 46.558704453441294 AUC: 0.628058369779245 F-Score: 0.5552873008208595 Confusion Matrix: [[985 261] [660 575]] ============================== ============ LightGbm =========== [LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines [LightGBM] [Info] Number of positive: 4967, number of negative: 4956 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000148 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 10 [LightGBM] [Info] Number of data points in the train set: 9923, number of used features: 5 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500554 -> initscore=0.002217 [LightGBM] [Info] Start training from score 0.002217 [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf Prediction Vector: [1. 0. 1. ... 0. 0. 0.] Accuracy: 62.918178153970175 Precision of event Happening: 67.79661016949152 Recall of event Happening: 48.582995951417004 AUC: 0.6285490086495408 F-Score: 0.5660377358490566 Confusion Matrix: [[961 285] [635 600]] ==============================
In [33]:
## 4. Classification based Algrithms Results with both CV and RFFS and addressing class imbalancing
res_rffs_cv = template.MachineLearningwithRFFS_CV_clf(SAll_bal,'test_successful', threshold=5,
algo_list=template.get_supported_algorithms_clf())
cert_degree2_Intermediate 17.925801
cert_degree2_A-Level 13.890728
studied_maths_Yes 12.359540
studied_maths_No 10.572414
cert_degree1_O-Level 5.970183
cert_degree1_Other boards 5.115042
discipline2_Science 4.391016
cert_degree1_Matriculation 3.401322
discipline2_Arts 3.286955
city_Others 2.780194
gender_Female 2.448693
gender_Male 2.088421
city_Karachi-2 1.685928
city_Karachi-3 1.677672
Province_No_Sindh 1.584745
city_Hyderabad 1.558019
city_Karachi-1 1.500453
cert_degree1_Aga Khan Board 1.468577
Province_Sindh 1.426194
cert_degree2_Aga Khan Board 1.134365
city_Lahore 1.041228
city_Islamabad 0.997018
cert_degree2_Other boards 0.958269
city_Peshawar 0.393137
city_Quetta 0.344085
dtype: float64
Selected Features =['cert_degree2_Intermediate', 'cert_degree2_A-Level', 'studied_maths_Yes', 'studied_maths_No', 'cert_degree1_O-Level', 'cert_degree1_Other boards']
(12404, 26)
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000128 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000175 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000444 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000394 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000224 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000218 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000411 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000413 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000522 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000120 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[11:23:32] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:33] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:33] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:34] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:34] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:35] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:35] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:36] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:36] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:23:36] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000135 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5581
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000134 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500045 -> initscore=0.000179
[LightGBM] [Info] Start training from score 0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000299 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000451 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11163, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.499955 -> initscore=-0.000179
[LightGBM] [Info] Start training from score -0.000179
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000307 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000206 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000103 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000216 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000524 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 5582
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.000397 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 11164, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.500000 -> initscore=0.000000
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
============ LogReg ===========
{'accuracy': 64.7187751813054, 'precision': 68.1284026873877, 'recall': 51.00994753519298, 'auc_val': 0.6471814451197341, 'f_score': 0.5738632122649998}
==============================
============ KNN ===========
{'accuracy': 48.24150658937901, 'precision': 34.032006566298506, 'recall': 10.141473170224923, 'auc_val': 0.48243104254324454, 'f_score': 0.14346931389175263}
==============================
============ GadientBoosting ===========
{'accuracy': 59.277962621195186, 'precision': 60.22295255654909, 'recall': 50.477585579969876, 'auc_val': 0.5927695963846035, 'f_score': 0.534969964303594}
==============================
============ AdaBoost ===========
{'accuracy': 64.76712328767124, 'precision': 68.13234746056325, 'recall': 51.15510882551556, 'auc_val': 0.6476657056776272, 'f_score': 0.5747670762477152}
==============================
============ SVM ===========
{'accuracy': 59.49567206467209, 'precision': 60.701079095888204, 'recall': 50.364708326840166, 'auc_val': 0.5949471456028259, 'f_score': 0.5354456286479738}
==============================
============ DecisionTree ===========
{'accuracy': 59.51180109693015, 'precision': 60.74379953334549, 'recall': 50.3485792945821, 'auc_val': 0.5951084359254065, 'f_score': 0.5353943150190741}
==============================
============ DeepNeuralNetwork ===========
{'accuracy': 60.00305424865483, 'precision': 61.19624897363066, 'recall': 50.7032881408758, 'auc_val': 0.6000141551088254, 'f_score': 0.5410048034571874}
==============================
============ RandomForest ===========
{'accuracy': 59.26989160666476, 'precision': 60.19173003374539, 'recall': 50.47761155264662, 'auc_val': 0.5926890810866968, 'f_score': 0.5350181784570495}
==============================
============ NaiveBayes ===========
{'accuracy': 61.50168958436224, 'precision': 60.54745670585887, 'recall': 54.895018440600495, 'auc_val': 0.6150116877045348, 'f_score': 0.5679304168024751}
==============================
============ MultiLayerPerceptron ===========
{'accuracy': 59.72101063138469, 'precision': 60.81089095940111, 'recall': 51.00958391771856, 'auc_val': 0.5972127421952106, 'f_score': 0.5402761709719227}
==============================
============ LightGbm ===========
{'accuracy': 59.60046528553975, 'precision': 60.81198659813466, 'recall': 51.02607656745104, 'auc_val': 0.5959985195574256, 'f_score': 0.5386670177586771}
==============================
============ XgBoost ===========
{'accuracy': 59.51180109693015, 'precision': 60.74379953334549, 'recall': 50.3485792945821, 'auc_val': 0.5951084359254065, 'f_score': 0.5353943150190741}
==============================
The results of aforementioned four scenarios indicate that the Random Forest is being successful in predicting the outcome variable.
The results indicate that:
• With the application of CV, the precision and accuracy increase in comparison to the benchmark category i.e., without CV and RFFS.
• The performance in terms all four criterions (precision, recall, AUC and accuracy) worsens a little bit with the application of features selection criterion in comparison to the benchmark category
• Selected Features are
'cert_degree2_Intermediate',
'studied_maths_No',
'cert_degree2_A-Level',
'cert_degree1_O-Level',
'studied_maths_Yes',
'cert_degree1_Other boards'
• With the application of both CV and RFFS criterions, almost all 11 algorithms underperform than the CV only based scenario but outperform the rest of two scenarios.
• The best performing algorithm seems to be the Random Forest. Hence, in next section we carry out detailed analysis of the problem through Random Forest to construct the maximum depth tree and small depth tree and find out the most important predictors of the test success probability.
• Furthermore, the comparison between focusing the class imbalance issue and without addressing the class imbalance issue show that the overall accuracy is high in case of without addressing the class imbalance issue. However, the precision and recall is significantly misleading and worst in the scenario of without addressing the class imbalance issue. The results with addressing the class imbalance issue are reported here and results without addressing the class imbalance issue are reported in Appendix-A for reference.
Detailed Analysis of Test Successful Classification Problem through Random Forest¶
In [34]:
from sklearn.tree import export_graphviz
import pydot
from IPython.display import Image
#Full Tree with Random Forest without specifying the Maximum Depth of Tree (Considering all 26 features)
template.RF_all(SAll_bal,label_col='test_successful')
Image(filename = 'tree_all.png')
accuracy = 67.88132860367624 precision= 66.77253478523896 recall= 71.18671396323766 auc_val= 0.6788132860367623 f_score= 0.6890900577493366
Out[34]:
Random Forest without specifying the Maximum Depth of Tree (Considering all 26 features)
In this analysis (function RF_all), we use all features (26 features)
• n_estimators=1000
• random_state = 42 and
• no bound on tree length
The results show that accuracy, precision and recall of event happening (test successful) is reasonable. The tree is saved as "tree_all.png" and can be viewed in the folder. The tree is very large and cannot be interpreted easily. Therefore, the Random Forest algorithm with allowing maximum depth of tree to 5 is being run.
In [35]:
#Random Forest with allowing Maximum Depth of Tree to 5(Considering all 26 features)
template.RF_all_small(SAll_bal,label_col='test_successful')
Image(filename = 'small_treeall.png')
accuracy = 66.22057400838439 precision= 70.24144869215291 recall= 56.28829409867785 auc_val= 0.6622057400838438 f_score= 0.6249552452559971
Out[35]:
Random Forest with allowing Maximum Depth of Tree to 5 (Considering all 26 features) In the above analysis (function RF_all_small), we use all features (26 features)
• n_estimators=1000
• random_state = 42 and
• max_depth = 5
The results show that accuracy is almost similar, and the precision improves. However, the recall falls of event happening (test successful). The tree is saved as "small_treeall.png" and can be viewed in folder. The tree indicates that important role in predicting the test outcome is played by:
- cert_degree1_O-Level (9th and 10th curriculum)
- cert_degree2_A-Level (11th and 12th class)
- cert_degree2_Intermediate (11th and 12th class)
- Studied Mathematics- Yes
- Studied Mathematics-NO
With these attributes the prediction accuracy is about 67% and precision is 70%.
In [36]:
template.RF_imp(SAll_bal,label_col='test_successful')
Variable: cert_degree2_A-Level Importance: 0.15 Variable: cert_degree2_Intermediate Importance: 0.14 Variable: studied_maths_No Importance: 0.11 Variable: studied_maths_Yes Importance: 0.11 Variable: cert_degree1_O-Level Importance: 0.09 Variable: cert_degree1_Other boards Importance: 0.05 Variable: discipline2_Arts Importance: 0.04 Variable: discipline2_Science Importance: 0.04 Variable: city_Others Importance: 0.03 Variable: cert_degree1_Matriculation Importance: 0.03 Variable: gender_Female Importance: 0.02 Variable: gender_Male Importance: 0.02 Variable: Province_Sindh Importance: 0.02 Variable: city_Hyderabad Importance: 0.02 Variable: city_Karachi-2 Importance: 0.02 Variable: city_Karachi-3 Importance: 0.02 Variable: Province_No_Sindh Importance: 0.02 Variable: city_Islamabad Importance: 0.01 Variable: city_Karachi-1 Importance: 0.01 Variable: city_Lahore Importance: 0.01 Variable: cert_degree1_Aga Khan Board Importance: 0.01 Variable: cert_degree2_Aga Khan Board Importance: 0.01 Variable: cert_degree2_Other boards Importance: 0.01 Variable: city_Peshawar Importance: 0.0 Variable: city_Quetta Importance: 0.0
In [37]:
#Full Tree with Random Forest without specifying the Maximum Depth of Tree (Considering the top 6 features from above)
template.RF_all_imp(SAll_bal,label_col='test_successful')
Image(filename = 'tree_all_imp.png')
df is all cleaned up.. accuracy = 65.26926797807158 precision= 64.29218231210383 recall= 68.68752015478877 auc_val= 0.6526926797807159 f_score= 0.6641721234798876
Out[37]:
In [38]:
#Random Forest with allowing Maximum Depth of Tree to 5(Considering the top 6 features from above)
template.RFall_small_imp(SAll_bal, label_col='test_successful')
Image(filename = 'small_treeall_imp.png')
df is all cleaned up.. accuracy = 65.26926797807158 precision= 64.29218231210383 recall= 68.68752015478877 auc_val= 0.6526926797807159 f_score= 0.6641721234798876
Out[38]:
Overall Results of Random Forest Algorithm The Random Forest results show that even by considering the first six features, which are:
- cert_degree2_A-Level
- cert_degree2_Intermediate
- studied_maths_No
- studied_maths_Yes
- cert_degree1_O-Level
- cert_degree1_Other boards
the overall accuracy, precision and recall of event happening (means being successful in test) only fall by about 2%. Overall, the analysis suggests that we should consider the Random Forest algorithm to predict the outcome of entry test with above six features with allowing the tree depth to maximum level of 5. The precision and accuracy improve by allowing the tree depth further but in that case the tree becomes very complicated and difficult to interpret. Following conclusion can be drawn from this analysis:
- Curriculum (both at 9 and 10th classes and 11th and 12th classes level) is the most important variable.
- Mathematics background plays a significant role in predicting the outcome of entry test. The city variable which was assumed to capture the socio-economic status does not seem significant. The socio-economic status is an important indicator, however, in our case the city variable may not be the right proxy to represent the socio-economic status of the candidate.
Regression-based Prediction of Mathematics Score¶
In [39]:
#Setting the sample for analysis of Mathematics Score
S_math= template.cleaningup(df1, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['Eng_Score','test_successful'])
df is all cleaned up..
To analyze the prediction of Mathematics and select the best candidate ML algorithm, we have applied ten different machine learning regression algorithms considering the following four different scenarios:
• Without focusing the cross validation (CV) and features selection.
• Only cross validation (CV: cross_valid_kfold) is considered
• Only feature selection (random forest based algorithm is used) criterion is considered
• Both cross validation (CV: cross_valid_kfold)
For features selection, we run the random forest based algorithm to extract the top candidate features.
Here, we report the following result of best scenarios:
• without focusing the cross validation (CV) and features selection
Results of remaining scenarios are reported in Appendix-B.
In [40]:
## **1. Regrssion based Algrithms Results without CV and RFFS**
#withoutCV and RFFS
results_without_cv_reg_rffs= template.run_algorithms_reg(S_math,'Math_Score')
============ LinearReg =========== R-Squared Value: 0.2402995380873324 Adjusted R-Squared: 0.2378523648431864 MAE: 12.087517667667266 RMSE: 14.922773615625953 ============================== ============ RidgeReg =========== R-Squared Value: 0.2401850772087837 Adjusted R-Squared: 0.23773753525932084 MAE: 12.089108874667195 RMSE: 14.923897749104757 ============================== ============ LassoReg =========== R-Squared Value: 0.24001385441764803 Adjusted R-Squared: 0.23756576091944437 MAE: 12.090235858494612 RMSE: 14.925579189813675 ============================== ============ RandomForestReg =========== R-Squared Value: 0.27017253878661396 Adjusted R-Squared: 0.26782159347926504 MAE: 11.77517456970479 RMSE: 14.626433995078552 ============================== ============ SupportVectorRegression =========== R-Squared Value: 0.268951664281338 Adjusted R-Squared: 0.26659678625106265 MAE: 11.66914216837546 RMSE: 14.638662624189385 ============================== ============ DecisionTreeReg =========== R-Squared Value: 0.26864135996448835 Adjusted R-Squared: 0.26628548237128036 MAE: 11.781973047061618 RMSE: 14.641769093391902 ============================== ============ DeepNeuralNetworkReg =========== R-Squared Value: 0.27504449284944676 Adjusted R-Squared: 0.2727092412480083 MAE: 11.76247885782811 RMSE: 14.5775329605308 ============================== ============ GradientBoostingReg =========== R-Squared Value: 0.277454956315514 Adjusted R-Squared: 0.27512746938185695 MAE: 11.740773258750009 RMSE: 14.553277769870318 ============================== ============ AdaBooostReg =========== R-Squared Value: 0.2136301185683922 Adjusted R-Squared: 0.21109703687327686 MAE: 12.483506419868462 RMSE: 15.182447605520007 ============================== ============ VotingReg =========== R-Squared Value: 0.2720484607746412 Adjusted R-Squared: 0.2697035582516888 MAE: 11.840735130351367 RMSE: 14.607624273100317 ==============================
The results indicate that: • With the application of CV and features selections based criterions, the R2 as well as RMSE worsen significantly. The benchmark scenario (without CV and RFFS) performs better in terms of both R2 as well as RMSE than the remaining three scenarios.
• The important features selected through Random Forest are 'cert_degree1_Matriculation' 'cert_degree2_A-Level' 'studied_maths_No' 'cert_degree1_Other boards' 'studied_maths_Yes'
• With the application of features selection criterion, the R2 falls about 2% and the RMSE increases about 3% in comparison to the benchmark scenario.
• The best performing algorithm seems to be the Gradient Boosting Regression and Random Forest which have R2 of about 28% and RMSE of 14%. The performance of all remaining algorithms is considerably worse.
Overall, the prediction performance of all ML algorithm is not encouraging. This may be due to the prediction of continuous variable (Mathematics score) only with the help of categorical variables. In my view, if we have the grades or percentage marks of the applicants in the last examination, i.e. A-Level / Intermediate, the prediction may improve significantly as it seems an important indicator of fitting and predicting the Mathematics score in the entry test.
The detailed results of all 4 scenarios are given in Appendix-B for reference.
Regression-based Prediction of English Score¶
In [41]:
#Setting the sample for analysis of English Score
S_eng= template.cleaningup(df1, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['Math_Score','test_successful'])
df is all cleaned up..
To analyze the prediction of English score in the entry test and select the best candidate ML algorithm, we have applied ten different machine learning regression algorithms considering the four different scenarios, which are
a. without focusing the cross validation (CV) and features selection.
b. Only cross validation (CV: cross_valid_kfold) is considered
c. Only feature selection (random forest based algorithm is used) criterion is considered
d. Both cross validation (CV: cross_valid_kfold)
For features selection, we run the random forest based algorithm to extract the top candidate features.
Here, we report the result of best scenarios which is
• without focusing the cross validation (CV) and features selection
and report the results of remaining scenarios in Appendix-C for reference.
In [42]:
## **1. Regrssion based Algrithms Results without CV and RFFS**
#without CV and RFFS
results_withoutCV_RFFS_Eng= template.run_algorithms_reg(S_eng,"Eng_Score")
============ LinearReg =========== R-Squared Value: 0.1619175121597779 Adjusted R-Squared: 0.15921785203917427 MAE: 12.901545685116218 RMSE: 16.143095230919723 ============================== ============ RidgeReg =========== R-Squared Value: 0.1615841133633732 Adjusted R-Squared: 0.15888337928710528 MAE: 12.899228458413164 RMSE: 16.14630586557721 ============================== ============ LassoReg =========== R-Squared Value: 0.1617707002308051 Adjusted R-Squared: 0.1590705671945688 MAE: 12.896473072133317 RMSE: 16.14450911034221 ============================== ============ RandomForestReg =========== R-Squared Value: 0.16485217397680463 Adjusted R-Squared: 0.16216196708973085 MAE: 12.860896996234725 RMSE: 16.11480680421095 ============================== ============ SupportVectorRegression =========== R-Squared Value: 0.16823407918232902 Adjusted R-Squared: 0.165554766204563 MAE: 12.811268299953362 RMSE: 16.082145500391903 ============================== ============ DecisionTreeReg =========== R-Squared Value: 0.15815266635889036 Adjusted R-Squared: 0.15544087878756863 MAE: 12.897246524804086 RMSE: 16.179313719587846 ============================== ============ DeepNeuralNetworkReg =========== R-Squared Value: 0.16994819220268298 Adjusted R-Squared: 0.16727440078470424 MAE: 12.853866563970735 RMSE: 16.06556581651595 ============================== ============ GradientBoostingReg =========== R-Squared Value: 0.17082286050072892 Adjusted R-Squared: 0.16815188659434033 MAE: 12.845414572738852 RMSE: 16.057099028799072 ============================== ============ AdaBooostReg =========== R-Squared Value: 0.12884926889875115 Adjusted R-Squared: 0.1260430882161676 MAE: 13.255650173366567 RMSE: 16.458493431867346 ============================== ============ VotingReg =========== R-Squared Value: 0.16908183045657077 Adjusted R-Squared: 0.16640524828435255 MAE: 12.880829460737855 RMSE: 16.073947800359743 ==============================
Results
The results indicate that
• With the application of CV and features selections based criterions, the R2 as well as RMSE worsen significantly. The benchmark scenario (without CV and RFFS) performs better in terms of both R2 as well as RMSE than the other three scenarios. However, the results are not as good as were for the Mathematics score analysis. The R2 falls from 27% to 16%.
• The important features selected through Random Forest are
'cert_degree1_Matriculation'
'cert_degree2_A-Level'
'studied_maths_No'
'cert_degree1_Other boards'
'studied_maths_Yes'
• with the application of features selection criterion, the R2 falls about 2% and the RMSE increases about 3% in comparison to the benchmark scenario.
• The best performing algorithm seems to be the Deep Neural Network based Regression model followed by Gradient Boosting which have R2 of about 17% and RMSE of 16%. The performance of all remaining algorithms is considerably worse.
Overall, the prediction performance of all ML algorithm is not encouraging. This may be due to the prediction of continuous variable (English score) only with the help of categorical variables. In my view, if we have the grades or percentage marks of the applicants in the last examination, i.e. A-Level / Intermediate, the prediction may improve significantly as it seems an important indicator of fitting and predicting the English score in the entry test.
The detailed results of all 4 scenarios are given in Appendix-C for reference.
Conclusion:¶
The purpose of this project was to gauge the prediction capability of different machine learnings models. We used the test scores of 38,931 candidates from year 2014 to 2020 and tried to predict:
1- Test successful (classification problem)
2- English score (regression problem)
3- Mathematics score (regression problem)
The results suggest that for classification problem, Random Forest model has better prediction power than any other algorithm. For regression problems, Gradient Boosting is more suitable for predicting mathematics score while for predicting English score, Deep Neural Network came out to be more suitable. The performance of Random Forest model is close to Deep Neural Network and Gradient Boosting models in the regression-based problems. Most important features for predicting the test performance of a candidates have been identified as curriculum of the candidate (A-Level/Intermediate) and whether the candidate has prior mathematics background.
Unavailability of prior academic scores in A-Level / Intermediate has been identified as a limitation of this study. The availability of this data may significantly improve the prediction power of the models used in the study.
Appendix-A: Classification-based Prediction of Success Without Addressing Class Imbalance¶
In [43]:
## **1. Classification based Algrithms Results without CV and RFFS**
#without REG, CV and RFFS and addressing class imbalancing
results_without_cvRFFS_cls= template.run_algorithms_clf(S_all,'test_successful')
============ LogReg =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.39707204314884 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.4999239312338354 F-Score: 0.0 Confusion Matrix: [[6572 1] [1214 0]] ============================== ============ KNN =========== Prediction Vector: [0 0 0 ... 0 0 1] Accuracy: 82.92025170155387 Precision of event Happening: 30.79470198675497 Recall of event Happening: 7.660626029654035 AUC: 0.5224047580198661 F-Score: 0.12269129287598944 Confusion Matrix: [[6364 209] [1121 93]] ============================== ============ GadientBoosting =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ AdaBoost =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.39707204314884 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.4999239312338354 F-Score: 0.0 Confusion Matrix: [[6572 1] [1214 0]] ============================== ============ SVM =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ DecisionTree =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.24296905098241 Precision of event Happening: 11.76470588235294 Recall of event Happening: 0.16474464579901155 AUC: 0.49968269173652585 F-Score: 0.0032493907392363935 Confusion Matrix: [[6558 15] [1212 2]] ============================== ============ DeepNeuralNetwork =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.3713882111211 Precision of event Happening: 20.0 Recall of event Happening: 0.08237232289950577 AUC: 0.5001075865498391 F-Score: 0.0016406890894175553 Confusion Matrix: [[6569 4] [1213 1]] ============================== ============ RandomForest =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.23012713496854 Precision of event Happening: 11.11111111111111 Recall of event Happening: 0.16474464579901155 AUC: 0.49960662297036124 F-Score: 0.0032467532467532465 Confusion Matrix: [[6557 16] [1212 2]] ============================== ============ NaiveBayes =========== Prediction Vector: [0 0 0 ... 1 0 1] Accuracy: 73.73828175163734 Precision of event Happening: 28.43798650752465 Recall of event Happening: 45.14003294892916 AUC: 0.6208013362036447 F-Score: 0.34893346068131165 Confusion Matrix: [[5194 1379] [ 666 548]] ============================== ============ MultiLayerPerceptron =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.3713882111211 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.49977179370150615 F-Score: 0.0 Confusion Matrix: [[6570 3] [1214 0]] ============================== ============ LightGbm =========== [LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines [LightGBM] [Info] Number of positive: 4988, number of negative: 26156 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001041 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 50 [LightGBM] [Info] Number of data points in the train set: 31144, number of used features: 25 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.160159 -> initscore=-1.657044 [LightGBM] [Info] Start training from score -1.657044 Prediction Vector: [0. 0. 0. ... 0. 0. 0.] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ XgBoost =========== [11:32:27] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior. Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ LightGbm =========== [LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines [LightGBM] [Info] Number of positive: 4988, number of negative: 26156 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.002246 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 50 [LightGBM] [Info] Number of data points in the train set: 31144, number of used features: 25 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.160159 -> initscore=-1.657044 [LightGBM] [Info] Start training from score -1.657044 Prediction Vector: [0. 0. 0. ... 0. 0. 0.] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ==============================
In [44]:
## **2. Classification based Algrithms Results with CV only**
#with CV only and with addressing class imbalancing
results_cv_clf = template.run_algorithms_cv_clf(S_all,'test_successful')
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.002927 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35037, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159289 -> initscore=-1.663530
[LightGBM] [Info] Start training from score -1.663530
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.003310 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.002180 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001393 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001705 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001267 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001245 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001197 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001575 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 29457
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001254 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159284 -> initscore=-1.663564
[LightGBM] [Info] Start training from score -1.663564
[11:42:50] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:42:52] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:42:55] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:42:58] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:43:00] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:43:03] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:43:05] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:43:08] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:43:10] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:43:13] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001194 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35037, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159289 -> initscore=-1.663530
[LightGBM] [Info] Start training from score -1.663530
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001262 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001266 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001257 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001203 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001245 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001259 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001226 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001214 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] Find whitespaces in feature_names, replace with underlines
[LightGBM] [Info] Number of positive: 5581, number of negative: 29457
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.001226 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 50
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 25
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159284 -> initscore=-1.663564
[LightGBM] [Info] Start training from score -1.663564
============ LogReg ===========
{'accuracy': 84.06411505195938, 'precision': 3.333333333333333, 'recall': 0.0322061191626409, 'auc_val': 0.5000999245768704, 'f_score': 0.0006379585326953748}
==============================
============ KNN ===========
{'accuracy': 80.8096650896853, 'precision': 28.387229626034646, 'recall': 16.28845254791959, 'auc_val': 0.5466266649053506, 'f_score': 0.18227350550654436}
==============================
============ GadientBoosting ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ AdaBoost ===========
{'accuracy': 84.06668310537489, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.4999847234952643, 'f_score': 0.0}
==============================
============ SVM ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ DecisionTree ===========
{'accuracy': 83.9947657358743, 'precision': 35.69047619047619, 'recall': 0.225728533582671, 'auc_val': 0.5004717342888438, 'f_score': 0.0044564262225685655}
==============================
============ DeepNeuralNetwork ===========
{'accuracy': 84.04613274111765, 'precision': 1.0, 'recall': 0.016129032258064516, 'auc_val': 0.4999278754450748, 'f_score': 0.0003174603174603174}
==============================
============ RandomForest ===========
{'accuracy': 84.0127460677383, 'precision': 41.76877289377289, 'recall': 0.3385538413588904, 'auc_val': 0.5010358608277249, 'f_score': 0.006684308944703306}
==============================
============ NaiveBayes ===========
{'accuracy': 75.10999751836195, 'precision': 33.292227412923914, 'recall': 42.94099007843748, 'auc_val': 0.6207380819187847, 'f_score': 0.3300230384443415}
==============================
============ MultiLayerPerceptron ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ LightGbm ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ XgBoost ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
In [45]:
## **3. Classification based Algrithms Results with RFFS only**
#with RFFS only and without addressing class imbalancing
res_rffs_clf = template.MachineLearningwithRFFS_clf(S_all,'test_successful', threshold=5,
algo_list=template.get_supported_algorithms_clf())
studied_maths_No 15.322639 cert_degree1_O-Level 14.020934 cert_degree2_A-Level 13.256182 cert_degree2_Intermediate 10.393659 studied_maths_Yes 10.209890 discipline2_Science 7.143059 discipline2_Arts 4.357948 cert_degree1_Other boards 3.465455 city_Others 2.065979 gender_Male 2.052568 gender_Female 2.052412 cert_degree1_Matriculation 1.799521 city_Karachi-2 1.624144 city_Karachi-3 1.566775 Province_No_Sindh 1.517475 city_Karachi-1 1.419193 cert_degree1_Aga Khan Board 1.183820 Province_Sindh 1.153186 city_Hyderabad 1.142658 city_Islamabad 1.097222 city_Lahore 0.931740 cert_degree2_Other boards 0.706921 cert_degree2_Aga Khan Board 0.682874 city_Quetta 0.435067 city_Peshawar 0.398681 dtype: float64 Selected Features =['studied_maths_No', 'cert_degree1_O-Level', 'cert_degree2_A-Level', 'cert_degree2_Intermediate', 'studied_maths_Yes', 'discipline2_Science'] (38931, 26) ============ LogReg =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ KNN =========== Prediction Vector: [0 0 0 ... 0 0 1] Accuracy: 83.16424810581738 Precision of event Happening: 40.831758034026464 Recall of event Happening: 17.792421746293247 AUC: 0.5651525849219424 F-Score: 0.24784853700516357 Confusion Matrix: [[6260 313] [ 998 216]] ============================== ============ GadientBoosting =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ AdaBoost =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ SVM =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ DecisionTree =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ DeepNeuralNetwork =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ RandomForest =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ NaiveBayes =========== Prediction Vector: [0 0 0 ... 1 0 1] Accuracy: 78.2714781045332 Precision of event Happening: 33.31005586592179 Recall of event Happening: 39.29159802306425 AUC: 0.623812318428116 F-Score: 0.36054421768707484 Confusion Matrix: [[5618 955] [ 737 477]] ============================== ============ MultiLayerPerceptron =========== Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ LightGbm =========== [LightGBM] [Info] Number of positive: 4988, number of negative: 26156 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000374 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 12 [LightGBM] [Info] Number of data points in the train set: 31144, number of used features: 6 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.160159 -> initscore=-1.657044 [LightGBM] [Info] Start training from score -1.657044 [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf Prediction Vector: [0. 0. 0. ... 0. 0. 0.] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ XgBoost =========== [11:43:51] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior. Prediction Vector: [0 0 0 ... 0 0 0] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ============================== ============ LightGbm =========== [LightGBM] [Info] Number of positive: 4988, number of negative: 26156 [LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000633 seconds. You can set `force_row_wise=true` to remove the overhead. And if memory is not enough, you can set `force_col_wise=true`. [LightGBM] [Info] Total Bins 12 [LightGBM] [Info] Number of data points in the train set: 31144, number of used features: 6 [LightGBM] [Info] [binary:BoostFromScore]: pavg=0.160159 -> initscore=-1.657044 [LightGBM] [Info] Start training from score -1.657044 [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf [LightGBM] [Warning] No further splits with positive gain, best gain: -inf Prediction Vector: [0. 0. 0. ... 0. 0. 0.] Accuracy: 84.40991395916271 Precision of event Happening: 0.0 Recall of event Happening: 0.0 AUC: 0.5 F-Score: 0.0 Confusion Matrix: [[6573 0] [1214 0]] ==============================
In [46]:
## **4. Classification based Algrithms Results with both CV and RFFS**
#with CV and RFFS and without addressing class imbalancing
res_rffs_cv = template.MachineLearningwithRFFS_CV_clf(S_all,'test_successful', threshold=5,
algo_list=template.get_supported_algorithms_clf())
cert_degree2_A-Level 18.147152
cert_degree2_Intermediate 12.266995
cert_degree1_O-Level 11.781272
studied_maths_Yes 9.712526
studied_maths_No 9.216365
discipline2_Arts 5.631796
cert_degree1_Other boards 4.815824
discipline2_Science 4.353571
cert_degree1_Matriculation 3.940280
gender_Male 2.272477
city_Others 1.960754
gender_Female 1.836935
cert_degree1_Aga Khan Board 1.641374
city_Karachi-2 1.517454
city_Karachi-3 1.359385
Province_No_Sindh 1.344822
city_Hyderabad 1.292613
city_Karachi-1 1.241562
Province_Sindh 1.096349
cert_degree2_Aga Khan Board 1.012247
cert_degree2_Other boards 0.985839
city_Lahore 0.883983
city_Islamabad 0.778803
city_Quetta 0.545144
city_Peshawar 0.364477
dtype: float64
Selected Features =['cert_degree2_A-Level', 'cert_degree2_Intermediate', 'cert_degree1_O-Level', 'studied_maths_Yes', 'studied_maths_No', 'discipline2_Arts']
(38931, 26)
[LightGBM] [Info] Number of positive: 5581, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000517 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35037, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159289 -> initscore=-1.663530
[LightGBM] [Info] Start training from score -1.663530
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000609 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000648 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000622 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000925 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000679 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.002472 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000641 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000604 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5581, number of negative: 29457
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.003770 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159284 -> initscore=-1.663564
[LightGBM] [Info] Start training from score -1.663564
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[11:49:21] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:22] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:23] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:24] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:25] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:27] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:28] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:29] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[11:49:31] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.4.0/src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
[LightGBM] [Info] Number of positive: 5581, number of negative: 29456
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.002676 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35037, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159289 -> initscore=-1.663530
[LightGBM] [Info] Start training from score -1.663530
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000620 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000641 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000590 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000639 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000555 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing col-wise multi-threading, the overhead of testing was 0.002402 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000704 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5582, number of negative: 29456
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000577 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159313 -> initscore=-1.663350
[LightGBM] [Info] Start training from score -1.663350
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Number of positive: 5581, number of negative: 29457
[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.000678 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 12
[LightGBM] [Info] Number of data points in the train set: 35038, number of used features: 6
[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.159284 -> initscore=-1.663564
[LightGBM] [Info] Start training from score -1.663564
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
============ LogReg ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ KNN ===========
{'accuracy': 80.71194317009274, 'precision': 6.656706766525927, 'recall': 8.85483870967742, 'auc_val': 0.5159210007588972, 'f_score': 0.07028938698372802}
==============================
============ GadientBoosting ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ AdaBoost ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ SVM ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ DecisionTree ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ DeepNeuralNetwork ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ RandomForest ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ NaiveBayes ===========
{'accuracy': 78.06652689806722, 'precision': 30.268572497520864, 'recall': 39.10435821515765, 'auc_val': 0.6227750755243064, 'f_score': 0.2938423775169174}
==============================
============ MultiLayerPerceptron ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ LightGbm ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
============ XgBoost ===========
{'accuracy': 84.0692511587904, 'precision': 0.0, 'recall': 0.0, 'auc_val': 0.5, 'f_score': 0.0}
==============================
Appendix-B: Regression-based Prediction of Mathematics Score¶
In [47]:
## **1. Regrssion based Algrithms Results without CV and RFFS**
#withoutCV and RFFS
results_without_cv_reg_rffs= template.run_algorithms_reg(S_math,'Math_Score')
============ LinearReg =========== R-Squared Value: 0.2402995380873324 Adjusted R-Squared: 0.2378523648431864 MAE: 12.087517667667266 RMSE: 14.922773615625953 ============================== ============ RidgeReg =========== R-Squared Value: 0.2401850772087837 Adjusted R-Squared: 0.23773753525932084 MAE: 12.089108874667195 RMSE: 14.923897749104757 ============================== ============ LassoReg =========== R-Squared Value: 0.24001385441764803 Adjusted R-Squared: 0.23756576091944437 MAE: 12.090235858494612 RMSE: 14.925579189813675 ============================== ============ RandomForestReg =========== R-Squared Value: 0.27043188371748983 Adjusted R-Squared: 0.2680817738209479 MAE: 11.77316662685523 RMSE: 14.623835005068985 ============================== ============ SupportVectorRegression =========== R-Squared Value: 0.268951664281338 Adjusted R-Squared: 0.26659678625106265 MAE: 11.66914216837546 RMSE: 14.638662624189385 ============================== ============ DecisionTreeReg =========== R-Squared Value: 0.2685229367140538 Adjusted R-Squared: 0.26616667765180047 MAE: 11.783285293832963 RMSE: 14.642954459646306 ============================== ============ DeepNeuralNetworkReg =========== R-Squared Value: 0.2754937278179337 Adjusted R-Squared: 0.27315992330761907 MAE: 11.73617591510554 RMSE: 14.57301561268801 ============================== ============ GradientBoostingReg =========== R-Squared Value: 0.277454956315514 Adjusted R-Squared: 0.27512746938185695 MAE: 11.740773258750009 RMSE: 14.553277769870318 ============================== ============ AdaBooostReg =========== R-Squared Value: 0.2136301185683922 Adjusted R-Squared: 0.21109703687327686 MAE: 12.483506419868462 RMSE: 15.182447605520007 ============================== ============ VotingReg =========== R-Squared Value: 0.2720482621849014 Adjusted R-Squared: 0.26970335902224485 MAE: 11.841620520438571 RMSE: 14.607626265625884 ==============================
In [48]:
## **2. Regrssion based Algrithms Results with CV only**
#with CV only
results_cv = template.run_algorithms_cv_reg(S_math,'Math_Score')
============ LinearReg ===========
{'r2': 0.119817631213965, 'r2_adjusted': 0.11412743639919756, 'mae': 12.132393954695818, 'rmse': 14.938626087459344}
==============================
============ RidgeReg ===========
{'r2': 0.11953099394379105, 'r2_adjusted': 0.11383894616609976, 'mae': 12.136241560133692, 'rmse': 14.941051235164633}
==============================
============ LassoReg ===========
{'r2': 0.11983724806078644, 'r2_adjusted': 0.11414717953812437, 'mae': 12.138256978376443, 'rmse': 14.939172528340602}
==============================
============ RandomForestReg ===========
{'r2': 0.1447826024563575, 'r2_adjusted': 0.1392538110408113, 'mae': 11.879836967105843, 'rmse': 14.716643427014642}
==============================
============ SupportVectorRegression ===========
{'r2': 0.1471053472837443, 'r2_adjusted': 0.14159156572092896, 'mae': 11.763965672436814, 'rmse': 14.696301186899149}
==============================
============ DecisionTreeReg ===========
{'r2': 0.14182108744639343, 'r2_adjusted': 0.13627315153188446, 'mae': 11.897582497203109, 'rmse': 14.74263312023071}
==============================
============ DeepNeuralNetworkReg ===========
{'r2': 0.1514410531403561, 'r2_adjusted': 0.1459552992478173, 'mae': 11.85841333098235, 'rmse': 14.655006087897926}
==============================
============ GradientBoostingReg ===========
{'r2': 0.1592226413369308, 'r2_adjusted': 0.1537871881628014, 'mae': 11.808568852128847, 'rmse': 14.589841839858485}
==============================
============ AdaBooostReg ===========
{'r2': 0.09673898549333208, 'r2_adjusted': 0.0908995829287338, 'mae': 12.487887144571532, 'rmse': 15.137655260738418}
==============================
============ VotingReg ===========
{'r2': 0.1538107999577179, 'r2_adjusted': 0.1483403621499137, 'mae': 11.907447375162198, 'rmse': 14.640497970008274}
==============================
In [49]:
## **3. Regrssion based Algrithms Results with RFFS only**
#with RFFS only
res_rffs = template.MachineLearningwithRFFS_reg(S_math,'Math_Score', threshold=5,
algo_list=template.get_supported_algorithms_reg())
cert_degree1_Matriculation 35.828490 cert_degree2_A-Level 14.021842 studied_maths_No 12.953509 cert_degree1_Other boards 10.789181 studied_maths_Yes 9.585195 discipline2_Arts 3.416340 discipline2_Science 2.746731 city_Others 1.665540 gender_Female 1.346942 gender_Male 1.047747 Province_No_Sindh 0.931605 city_Hyderabad 0.671787 Province_Sindh 0.587022 city_Karachi-3 0.584834 cert_degree2_Intermediate 0.550001 city_Karachi-2 0.539999 city_Karachi-1 0.487054 city_Islamabad 0.430850 cert_degree2_Aga Khan Board 0.354418 city_Lahore 0.321752 cert_degree2_Other boards 0.291224 cert_degree1_Aga Khan Board 0.269776 cert_degree1_O-Level 0.213173 city_Quetta 0.185110 city_Peshawar 0.179879 dtype: float64 Selected Features =['cert_degree1_Matriculation', 'cert_degree2_A-Level', 'studied_maths_No', 'cert_degree1_Other boards', 'studied_maths_Yes'] (38931, 26) ============ LinearReg =========== R-Squared Value: 0.23012710566737582 Adjusted R-Squared: 0.22963239233083 MAE: 12.176264294156125 RMSE: 15.022349785396068 ============================== ============ RidgeReg =========== R-Squared Value: 0.23014280897869424 Adjusted R-Squared: 0.2296481057329537 MAE: 12.176604186819606 RMSE: 15.022196577093702 ============================== ============ LassoReg =========== R-Squared Value: 0.23014235133500172 Adjusted R-Squared: 0.22964764779518354 MAE: 12.177938412274251 RMSE: 15.022201042085761 ============================== ============ RandomForestReg =========== R-Squared Value: 0.25317877049628046 Adjusted R-Squared: 0.25269886995039703 MAE: 11.973952473768014 RMSE: 14.795739721616645 ============================== ============ SupportVectorRegression =========== R-Squared Value: 0.244454941760189 Adjusted R-Squared: 0.24396943536111448 MAE: 11.910173361309912 RMSE: 14.881905418379494 ============================== ============ DecisionTreeReg =========== R-Squared Value: 0.2531613014280315 Adjusted R-Squared: 0.2526813896566834 MAE: 11.973962411497725 RMSE: 14.795912765891048 ============================== ============ DeepNeuralNetworkReg =========== R-Squared Value: 0.25243704792757093 Adjusted R-Squared: 0.2519566707574947 MAE: 11.987808361700452 RMSE: 14.80308526171276 ============================== ============ GradientBoostingReg =========== R-Squared Value: 0.2524005378117753 Adjusted R-Squared: 0.25192013718063 MAE: 11.980804753126675 RMSE: 14.80344674013863 ============================== ============ AdaBooostReg =========== R-Squared Value: 0.1924925334288794 Adjusted R-Squared: 0.19197363645768606 MAE: 12.65885810270255 RMSE: 15.38514624160653 ============================== ============ VotingReg =========== R-Squared Value: 0.2480773023259889 Adjusted R-Squared: 0.24759412362294686 MAE: 12.081574697630113 RMSE: 14.84618789542 ==============================
In [50]:
## **4. Regrssion based Algrithms Results with both CV and RFFS**
#with CV and RFFS both
res_rffs_cv = template.MachineLearningwithRFFS_CV_reg(S_math,"Math_Score", threshold=5,
algo_list=template.get_supported_algorithms_reg())
cert_degree1_Matriculation 36.577758
cert_degree2_A-Level 14.112391
studied_maths_No 12.888915
cert_degree1_Other boards 10.403025
studied_maths_Yes 9.328163
discipline2_Arts 3.172560
discipline2_Science 2.689242
city_Others 1.788182
gender_Male 1.194272
gender_Female 1.157825
Province_No_Sindh 0.836262
Province_Sindh 0.735462
city_Karachi-3 0.692371
city_Hyderabad 0.617070
cert_degree2_Intermediate 0.608585
city_Karachi-1 0.549428
city_Karachi-2 0.526237
cert_degree2_Aga Khan Board 0.400363
city_Islamabad 0.377329
cert_degree2_Other boards 0.293337
cert_degree1_Aga Khan Board 0.237166
city_Lahore 0.226833
cert_degree1_O-Level 0.207900
city_Quetta 0.195749
city_Peshawar 0.183576
dtype: float64
Selected Features =['cert_degree1_Matriculation', 'cert_degree2_A-Level', 'studied_maths_No', 'cert_degree1_Other boards', 'studied_maths_Yes']
(38931, 26)
============ LinearReg ===========
{'r2': 0.11067959704215165, 'r2_adjusted': 0.10953565872107607, 'mae': 12.21323892447696, 'rmse': 15.016267472367272}
==============================
============ RidgeReg ===========
{'r2': 0.11008827402572133, 'r2_adjusted': 0.10894357503236995, 'mae': 12.213222262830923, 'rmse': 15.018987497713406}
==============================
============ LassoReg ===========
{'r2': 0.11018591502122473, 'r2_adjusted': 0.10904134162207571, 'mae': 12.21520411386183, 'rmse': 15.018690331334108}
==============================
============ RandomForestReg ===========
{'r2': 0.13679350797882012, 'r2_adjusted': 0.13568315953425755, 'mae': 11.990635725548577, 'rmse': 14.78223165990487}
==============================
============ SupportVectorRegression ===========
{'r2': 0.12650972139656141, 'r2_adjusted': 0.12538614553826272, 'mae': 11.934004905862933, 'rmse': 14.870419571334732}
==============================
============ DecisionTreeReg ===========
{'r2': 0.13691925208786704, 'r2_adjusted': 0.13580906537732754, 'mae': 11.990011609126432, 'rmse': 14.7811458714656}
==============================
============ DeepNeuralNetworkReg ===========
{'r2': 0.13578207100251669, 'r2_adjusted': 0.13467042151630473, 'mae': 11.994658406709913, 'rmse': 14.790540065967093}
==============================
============ GradientBoostingReg ===========
{'r2': 0.13633711935876505, 'r2_adjusted': 0.13522618395552413, 'mae': 11.994752754275172, 'rmse': 14.786685390573524}
==============================
============ AdaBooostReg ===========
{'r2': 0.08822500471289468, 'r2_adjusted': 0.08705218183362287, 'mae': 12.551444836285475, 'rmse': 15.214888959937934}
==============================
============ VotingReg ===========
{'r2': 0.1348119705653919, 'r2_adjusted': 0.1336990729799059, 'mae': 12.068629251590659, 'rmse': 14.802836562031198}
==============================
Appendix-C: Regression-based Prediction of English Score¶
In [51]:
## **1. Regrssion based Algrithms Results without CV and RFFS**
#without CV and RFFS
results_withoutCV_RFFS_Eng= template.run_algorithms_reg(S_eng,"Eng_Score")
============ LinearReg =========== R-Squared Value: 0.1619175121597779 Adjusted R-Squared: 0.15921785203917427 MAE: 12.901545685116218 RMSE: 16.143095230919723 ============================== ============ RidgeReg =========== R-Squared Value: 0.1615841133633732 Adjusted R-Squared: 0.15888337928710528 MAE: 12.899228458413164 RMSE: 16.14630586557721 ============================== ============ LassoReg =========== R-Squared Value: 0.1617707002308051 Adjusted R-Squared: 0.1590705671945688 MAE: 12.896473072133317 RMSE: 16.14450911034221 ============================== ============ RandomForestReg =========== R-Squared Value: 0.16457927960802743 Adjusted R-Squared: 0.16188819366423157 MAE: 12.86563149671125 RMSE: 16.117439440438524 ============================== ============ SupportVectorRegression =========== R-Squared Value: 0.16823407918232902 Adjusted R-Squared: 0.165554766204563 MAE: 12.811268299953362 RMSE: 16.082145500391903 ============================== ============ DecisionTreeReg =========== R-Squared Value: 0.15685087299366318 Adjusted R-Squared: 0.15413489204080166 MAE: 12.904691296153171 RMSE: 16.191818354532597 ============================== ============ DeepNeuralNetworkReg =========== R-Squared Value: 0.17164375616477645 Adjusted R-Squared: 0.16897542655572084 MAE: 12.818936685455407 RMSE: 16.049148696869736 ============================== ============ GradientBoostingReg =========== R-Squared Value: 0.17082286050072892 Adjusted R-Squared: 0.16815188659434033 MAE: 12.84541457273885 RMSE: 16.057099028799072 ============================== ============ AdaBooostReg =========== R-Squared Value: 0.12884926889875115 Adjusted R-Squared: 0.1260430882161676 MAE: 13.255650173366567 RMSE: 16.458493431867346 ============================== ============ VotingReg =========== R-Squared Value: 0.16889705100936792 Adjusted R-Squared: 0.16621987361924218 MAE: 12.881908201772177 RMSE: 16.075734961954947 ==============================
In [52]:
## **2. Regrssion based Algrithms Results with CV only**
#with CV only
results_cv = template.run_algorithms_cv_reg(S_eng,"Eng_Score")
============ LinearReg ===========
{'r2': -0.06374830432619602, 'r2_adjusted': -0.07062521875831743, 'mae': 13.409836269070501, 'rmse': 16.362843753419433}
==============================
============ RidgeReg ===========
{'r2': -0.06250295856305624, 'r2_adjusted': -0.06937182162892754, 'mae': 13.397963645323424, 'rmse': 16.352274604557977}
==============================
============ LassoReg ===========
{'r2': -0.06232480409897763, 'r2_adjusted': -0.06919251466851953, 'mae': 13.401346580274708, 'rmse': 16.353583194602752}
==============================
============ RandomForestReg ===========
{'r2': -0.07296821264810238, 'r2_adjusted': -0.07990472823592908, 'mae': 13.441739909788396, 'rmse': 16.44533313249342}
==============================
============ SupportVectorRegression ===========
{'r2': -0.07478343530664137, 'r2_adjusted': -0.08173169022114213, 'mae': 13.451444364384566, 'rmse': 16.450021439277187}
==============================
============ DecisionTreeReg ===========
{'r2': -0.08412468446109278, 'r2_adjusted': -0.0911333177454846, 'mae': 13.506112153008491, 'rmse': 16.539302129769684}
==============================
============ DeepNeuralNetworkReg ===========
{'r2': -0.06485211897436063, 'r2_adjusted': -0.07173616989778976, 'mae': 13.40515312002658, 'rmse': 16.379945976186377}
==============================
============ GradientBoostingReg ===========
{'r2': -0.05718594615583376, 'r2_adjusted': -0.06402043411460799, 'mae': 13.378779806336771, 'rmse': 16.320852125331772}
==============================
============ AdaBooostReg ===========
{'r2': -0.07938407612566692, 'r2_adjusted': -0.08636205715270387, 'mae': 13.583447222754819, 'rmse': 16.49849737460031}
==============================
============ VotingReg ===========
{'r2': -0.05726182376812455, 'r2_adjusted': -0.06409680010269361, 'mae': 13.38210139630279, 'rmse': 16.32419860152027}
==============================
In [53]:
## **3. Regrssion based Algrithms Results with RFFS only**
##with RFFS only English
res_rffs = template.MachineLearningwithRFFS_reg(S_eng,"Eng_Score", threshold=5,
algo_list=template.get_supported_algorithms_reg())
cert_degree1_Matriculation 48.174467 cert_degree2_A-Level 7.748784 cert_degree1_Other boards 7.344466 discipline2_Science 6.333579 city_Others 3.842816 gender_Female 3.755714 studied_maths_Yes 2.269141 Province_Sindh 1.770020 gender_Male 1.684217 studied_maths_No 1.616758 Province_No_Sindh 1.615782 cert_degree2_Intermediate 1.586827 city_Karachi-1 1.383760 discipline2_Arts 1.374081 city_Karachi-2 1.249481 city_Hyderabad 1.216204 city_Islamabad 1.215400 city_Karachi-3 1.154083 cert_degree1_O-Level 0.965658 cert_degree2_Other boards 0.842930 cert_degree1_Aga Khan Board 0.841214 city_Lahore 0.751161 cert_degree2_Aga Khan Board 0.533253 city_Quetta 0.424958 city_Peshawar 0.305246 dtype: float64 Selected Features =['cert_degree1_Matriculation', 'cert_degree2_A-Level', 'cert_degree1_Other boards', 'discipline2_Science'] (38931, 26) ============ LinearReg =========== R-Squared Value: 0.1395245039419114 Adjusted R-Squared: 0.13908221378716557 MAE: 13.10569502865673 RMSE: 16.357339944616015 ============================== ============ RidgeReg =========== R-Squared Value: 0.13952450694496288 Adjusted R-Squared: 0.13908221679176058 MAE: 13.105695255109184 RMSE: 16.357339916072533 ============================== ============ LassoReg =========== R-Squared Value: 0.1395200797119257 Adjusted R-Squared: 0.13907778728309605 MAE: 13.106836071250843 RMSE: 16.3573819960987 ============================== ============ RandomForestReg =========== R-Squared Value: 0.1425370452589877 Adjusted R-Squared: 0.14209630357060887 MAE: 13.096607030437154 RMSE: 16.328681157476854 ============================== ============ SupportVectorRegression =========== R-Squared Value: 0.1378737120259267 Adjusted R-Squared: 0.13743057335310527 MAE: 13.06988492833393 RMSE: 16.373022916538 ============================== ============ DecisionTreeReg =========== R-Squared Value: 0.14256893256204317 Adjusted R-Squared: 0.14212820726395115 MAE: 13.096891683146907 RMSE: 16.32837753943568 ============================== ============ DeepNeuralNetworkReg =========== R-Squared Value: 0.14349223949382894 Adjusted R-Squared: 0.14305198878166947 MAE: 13.086412995396548 RMSE: 16.319583733423645 ============================== ============ GradientBoostingReg =========== R-Squared Value: 0.14293973139502347 Adjusted R-Squared: 0.14249919669000932 MAE: 13.09680566928021 RMSE: 16.32484652785684 ============================== ============ AdaBooostReg =========== R-Squared Value: 0.1321149414578845 Adjusted R-Squared: 0.13166884273851054 MAE: 13.23053752703005 RMSE: 16.427615585903883 ============================== ============ VotingReg =========== R-Squared Value: 0.1422489392266164 Adjusted R-Squared: 0.14180804944981185 MAE: 13.115786650644218 RMSE: 16.33142413104539 ==============================
In [54]:
## **4. Regrssion based Algrithms Results with both CV and RFFS**
#with CV and RFFS
res_rffs_cv = template.MachineLearningwithRFFS_CV_reg(S_eng,"Eng_Score", threshold=5,
algo_list=template.get_supported_algorithms_reg())
cert_degree1_Matriculation 48.119480
cert_degree2_A-Level 8.180574
cert_degree1_Other boards 6.544489
discipline2_Science 6.486983
city_Others 4.441244
gender_Female 3.998379
studied_maths_Yes 2.584433
Province_No_Sindh 1.840755
cert_degree1_O-Level 1.755020
cert_degree2_Intermediate 1.669336
city_Karachi-1 1.431928
Province_Sindh 1.318947
studied_maths_No 1.285581
city_Karachi-2 1.212501
gender_Male 1.155817
city_Hyderabad 1.126682
city_Islamabad 1.081441
city_Karachi-3 1.070356
cert_degree2_Other boards 1.022953
discipline2_Arts 0.832759
cert_degree1_Aga Khan Board 0.779103
city_Lahore 0.773639
cert_degree2_Aga Khan Board 0.493797
city_Quetta 0.455373
city_Peshawar 0.338430
dtype: float64
Selected Features =['cert_degree1_Matriculation', 'cert_degree2_A-Level', 'cert_degree1_Other boards', 'discipline2_Science']
(38931, 26)
============ LinearReg ===========
{'r2': -0.08342737592367298, 'r2_adjusted': -0.0845419840032135, 'mae': 13.568115501842925, 'rmse': 16.52474704353112}
==============================
============ RidgeReg ===========
{'r2': -0.08342730934561178, 'r2_adjusted': -0.08454191735665532, 'mae': 13.568115315961322, 'rmse': 16.524746619578906}
==============================
============ LassoReg ===========
{'r2': -0.08339329434585026, 'r2_adjusted': -0.08450786734149843, 'mae': 13.569232263207661, 'rmse': 16.525068463462333}
==============================
============ RandomForestReg ===========
{'r2': -0.08222382142207117, 'r2_adjusted': -0.08333719133734113, 'mae': 13.56436537390814, 'rmse': 16.519348738346736}
==============================
============ SupportVectorRegression ===========
{'r2': -0.1025213812593491, 'r2_adjusted': -0.10365563376743532, 'mae': 13.65869924303947, 'rmse': 16.66127814177455}
==============================
============ DecisionTreeReg ===========
{'r2': -0.08220097708436315, 'r2_adjusted': -0.08331432350590594, 'mae': 13.563859444669683, 'rmse': 16.5187696290696}
==============================
============ DeepNeuralNetworkReg ===========
{'r2': -0.08221029177649627, 'r2_adjusted': -0.08332364834542844, 'mae': 13.565084542760095, 'rmse': 16.516366330709296}
==============================
============ GradientBoostingReg ===========
{'r2': -0.0820968837946062, 'r2_adjusted': -0.08321012310469468, 'mae': 13.563312356517397, 'rmse': 16.518145793289637}
==============================
============ AdaBooostReg ===========
{'r2': -0.07739814170893856, 'r2_adjusted': -0.07850654568770091, 'mae': 13.5673623343632, 'rmse': 16.481546957400717}
==============================
============ VotingReg ===========
{'r2': -0.07805544918865742, 'r2_adjusted': -0.07916453041533844, 'mae': 13.547408349605249, 'rmse': 16.488253344115112}
==============================
Appendix-D: ANOVA Analysis of English and Mathematics Scores Comparisons Across all Categorical Variables¶
In [55]:
#seting data for plots
FILE_NAME = "FinalData_testing.csv"
df3 = template.load_data(FILE_NAME)
df3 = template.cleaningup(df3, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['term_name','date_of_birth', 'place_of_birth',
'postal_address', 'city1', 'countryname','seat_no',
'medium','test_center','interview_successful',
'discipline_Mat'])
#df3['Year'].values.astype(str)
#,to_date=['Year'])
df is all cleaned up..
In [56]:
df4=template.cleaningup(df3, to_numeric=[], cols_to_interpolate=[],
cols_to_delete=['Year'])
template.ANOVA_analysis(df4)
df is all cleaned up.. ============+++++============+++++============ Analysis of Columns Eng_Score by gender ============+++++============+++++============ Anova => - gender
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.014502e+07 | 1.0 | 97391.215552 | 0.000000e+00 |
| C(Q("gender")) | 8.660059e+04 | 1.0 | 279.785440 | 1.384166e-62 |
| Residual | 1.204950e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by gender ============+++++============+++++============ Anova => - gender
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.592565e+07 | 1.0 | 55044.080352 | 0.000000e+00 |
| C(Q("gender")) | 1.049090e+05 | 1.0 | 362.598437 | 1.779221e-80 |
| Residual | 1.126315e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by Province ============+++++============+++++============ Anova => - Province
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.455305e+05 | 1.0 | 468.910464 | 2.258974e-103 |
| C(Q("Province")) | 5.600229e+04 | 7.0 | 25.777669 | 1.912233e-35 |
| Residual | 1.208010e+07 | 38923.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by Province ============+++++============+++++============ Anova => - Province
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 9.095705e+04 | 1.0 | 311.841085 | 1.620726e-69 |
| C(Q("Province")) | 1.509261e+04 | 7.0 | 7.392022 | 6.650152e-09 |
| Residual | 1.135297e+07 | 38923.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by city ============+++++============+++++============ Anova => - city
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 2.877417e+06 | 1.0 | 9450.781736 | 0.000000e+00 |
| C(Q("city")) | 2.857754e+05 | 8.0 | 117.327487 | 6.429085e-195 |
| Residual | 1.185032e+07 | 38922.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by city ============+++++============+++++============ Anova => - city
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.986099e+06 | 1.0 | 6808.878721 | 0.000000e+00 |
| C(Q("city")) | 1.480408e+04 | 8.0 | 6.344046 | 2.966714e-08 |
| Residual | 1.135325e+07 | 38922.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by test_successful ============+++++============+++++============ Anova => - test_successful
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 5.780768e+07 | 1.0 | 206774.733225 | 0.0 |
| C(Q("test_successful")) | 1.252782e+06 | 1.0 | 4481.128247 | 0.0 |
| Residual | 1.088332e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by test_successful ============+++++============+++++============ Anova => - test_successful
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.465520e+07 | 1.0 | 151196.208931 | 0.0 |
| C(Q("test_successful")) | 2.445267e+06 | 1.0 | 10668.385871 | 0.0 |
| Residual | 8.922792e+06 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by cert_degree1 ============+++++============+++++============ Anova => - cert_degree1
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 2.058587e+06 | 1.0 | 7410.358798 | 0.0 |
| C(Q("cert_degree1")) | 1.322233e+06 | 3.0 | 1586.560912 | 0.0 |
| Residual | 1.081387e+07 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by cert_degree1 ============+++++============+++++============ Anova => - cert_degree1
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.135856e+06 | 1.0 | 4553.895797 | 0.0 |
| C(Q("cert_degree1")) | 1.658690e+06 | 3.0 | 2216.683801 | 0.0 |
| Residual | 9.709370e+06 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by cert_degree2 ============+++++============+++++============ Anova => - cert_degree2
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 4.639299e+07 | 1.0 | 160595.265526 | 0.0 |
| C(Q("cert_degree2")) | 8.908121e+05 | 3.0 | 1027.886759 | 0.0 |
| Residual | 1.124529e+07 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by cert_degree2 ============+++++============+++++============ Anova => - cert_degree2
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.115302e+07 | 1.0 | 113883.053281 | 0.0 |
| C(Q("cert_degree2")) | 7.194733e+05 | 3.0 | 876.702924 | 0.0 |
| Residual | 1.064859e+07 | 38927.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by studied_maths ============+++++============+++++============ Anova => - studied_maths
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.176566e+07 | 1.0 | 102828.769799 | 0.000000e+00 |
| C(Q("studied_maths")) | 1.102276e+05 | 1.0 | 356.818103 | 3.143113e-79 |
| Residual | 1.202587e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by studied_maths ============+++++============+++++============ Anova => - studied_maths
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 1.805425e+07 | 1.0 | 61907.726386 | 0.000000e+00 |
| C(Q("studied_maths")) | 1.513008e+04 | 1.0 | 51.880802 | 6.004120e-13 |
| Residual | 1.135293e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by discipline2 ============+++++============+++++============ Anova => - discipline2
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.669140e+07 | 1.0 | 117762.228397 | 0.000000 |
| C(Q("discipline2")) | 6.917344e+03 | 1.0 | 22.201438 | 0.000002 |
| Residual | 1.212918e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by discipline2 ============+++++============+++++============ Anova => - discipline2
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 2.288393e+07 | 1.0 | 78421.409334 | 0.000000e+00 |
| C(Q("discipline2")) | 8.296452e+03 | 1.0 | 28.431279 | 9.762422e-08 |
| Residual | 1.135976e+07 | 38929.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Eng_Score by program ============+++++============+++++============ Anova => - program
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 3.683319e+07 | 1.0 | 120462.186487 | 0.000000e+00 |
| C(Q("program")) | 2.341742e+05 | 5.0 | 153.172400 | 1.122473e-161 |
| Residual | 1.190192e+07 | 38925.0 | NaN | NaN |
============+++++============+++++============ Analysis of Columns Math_Score by program ============+++++============+++++============ Anova => - program
| sum_sq | df | F | PR(>F) | |
|---|---|---|---|---|
| Intercept | 2.306720e+07 | 1.0 | 81525.351325 | 0.000000e+00 |
| C(Q("program")) | 3.544185e+05 | 5.0 | 250.520992 | 2.173036e-264 |
| Residual | 1.101364e+07 | 38925.0 | NaN | NaN |
The results of ANOVA analysis indicate that the English and Mathematics is significantly affected by the categorical columns, which are:
{'gender', 'Province', 'city', 'program','test_successful', 'cert_degree2','discipline2','studied_maths'}
The average English and Mathematics scores significantly vary across the categories in each of the above variables, which is also evident in Appendix-E.
Appendix-E: Average English and Mathematics Scores Comparisons Across all Categorical Variables¶
In [57]:
template.Avg_by_cat1(df3,cat_list=['Year','gender', 'Province', 'city', 'program','test_successful', 'cert_degree2',
'discipline2','studied_maths'])
Analysis of Average English and Mathematics Score by Year
Analysis of Average English and Mathematics Score by gender
Analysis of Average English and Mathematics Score by Province
Analysis of Average English and Mathematics Score by city
Analysis of Average English and Mathematics Score by program
Analysis of Average English and Mathematics Score by test_successful
Analysis of Average English and Mathematics Score by cert_degree2
Analysis of Average English and Mathematics Score by discipline2
Analysis of Average English and Mathematics Score by studied_maths
Average English and Mathematics Score Analysis across Categorical Variables
• The average English score is higher for females’ candidates as compared to the male candidates, however, the Mathematics score of male applicants is significantly higher than the female.
• The average English score for backward areas such as Balochistan and KPK (except Gilgit) is lower than the average score of Islamabad, Punjab and Sindh students. The students coming from abroad has the highest English score. For Mathematics, the average score of Balochistan and KPK students is higher or almost similar to the students coming from Sindh, Punjab and AJK. The students from Gilgit are performing very well in both scores as compared to other far flunged areas.
• The student from Karachi 1 (which is DHA and Clifton area of Karachi), Lahore and Islamabad are equally well in English and have the highest score, while in Mathematics are not worse than the other cities.
• The BBA, ACF, SSLA and Economics students has the higher average score in English than the CS and BSEM students, while the latter group outperform the former in Mathematics section of the test.
• The A-level students have significant higher score than the students studying other curriculums both in English and Mathematics. The A-Level and other international boards students have almost same score in English, however the A-level students outperform them in the Mathematics part of the test.
Appendix-F: Percentage Counts Analysis of all Categorical Variables¶
In [58]:
template.Count_Per1(df3, label_col=['Year', 'gender','program', 'Province','city', 'cert_degree2',
'discipline2', 'test_successful','studied_maths'],
cat_list=['Year', 'gender','program', 'Province','city', 'cert_degree2','discipline2',
'test_successful','studied_maths'])
Analsis of Year by gender
Analsis of Year by program
Analsis of Year by Province
Analsis of Year by city
Analsis of Year by cert_degree2
Analsis of Year by discipline2
Analsis of Year by test_successful
Analsis of Year by studied_maths
Analsis of gender by Year
Analsis of gender by program
Analsis of gender by Province
Analsis of gender by city
Analsis of gender by cert_degree2
Analsis of gender by discipline2
Analsis of gender by test_successful
Analsis of gender by studied_maths
Analsis of program by Year
Analsis of program by gender
Analsis of program by Province
Analsis of program by city
Analsis of program by cert_degree2
Analsis of program by discipline2
Analsis of program by test_successful
Analsis of program by studied_maths
Analsis of Province by Year
Analsis of Province by gender
Analsis of Province by program
Analsis of Province by city
Analsis of Province by cert_degree2
Analsis of Province by discipline2
Analsis of Province by test_successful
Analsis of Province by studied_maths
Analsis of city by Year
Analsis of city by gender
Analsis of city by program
Analsis of city by Province
Analsis of city by cert_degree2
Analsis of city by discipline2
Analsis of city by test_successful
Analsis of city by studied_maths
Analsis of cert_degree2 by Year
Analsis of cert_degree2 by gender
Analsis of cert_degree2 by program
Analsis of cert_degree2 by Province
Analsis of cert_degree2 by city
Analsis of cert_degree2 by discipline2
Analsis of cert_degree2 by test_successful
Analsis of cert_degree2 by studied_maths
Analsis of discipline2 by Year
Analsis of discipline2 by gender
Analsis of discipline2 by program
Analsis of discipline2 by Province
Analsis of discipline2 by city
Analsis of discipline2 by cert_degree2
Analsis of discipline2 by test_successful
Analsis of discipline2 by studied_maths
Analsis of test_successful by Year
Analsis of test_successful by gender
Analsis of test_successful by program
Analsis of test_successful by Province
Analsis of test_successful by city
Analsis of test_successful by cert_degree2
Analsis of test_successful by discipline2
Analsis of test_successful by studied_maths
Analsis of studied_maths by Year
Analsis of studied_maths by gender
Analsis of studied_maths by program
Analsis of studied_maths by Province
Analsis of studied_maths by city
Analsis of studied_maths by cert_degree2
Analsis of studied_maths by discipline2
Analsis of studied_maths by test_successful
Appendix-F
Percentage Counts of Categorical Variables Analysis
The analysis of data across year shows that
• percentage of male and female applicants is consistent over the years. On an average 35.85% of the applicants were female whereas 64.14% were male
• BBA as the most prominent degree as over the years it had most number of applicants. BSECO has the lowest percentage of applicants as the degree program was introduced later.
• most number of applicants were populated by those residing in the Sindh region
• Karachi (the capital city of the province Sindh) was the most popular area from where applications came in.
• there were two almost equal divisions in the applicants. On an average 49.78% of the applicants had A-Level background whereas 46.98% of the applicants did intermediate.
• majority of the applicants had science background.
• reflects that over the years’ success rate has decreased by approximately 8% and by similar percentage the number of unsuccessful candidates have increased.
• It is clearly evident that those applying at the institute had a mathematical background as they studied the subject before joining.
Analysis of Column gender Gender based analysis reveals that
• Percentage of female applicants in the year 2015 almost doubled to what it was in 2014. Over the years the percentage of both male and female applicants have gradually increased, however, both genders witnessed a slight dip in the year 2017.
• the degree program of BBA & SSLA is dominated by female applicants. Comparatively, the degrees such as BSACF & BSCS was tilted towards the male.
• male applicants have taken the lead in all provinces except for the Foreign & Sindh state. The Foreign & Sindh region is the only anomaly where percentage of female applicants is greater than male applicants.
• Sindh’s popular city of Karachi had more female applicants than male applicants. In other provinces it was the opposite as the percentage of male applicants was greater.
• Most percentage of the female applicants had the background of A-Level education system. Whereas the male population was almost equally distributed A-Level and Intermediate Board.
• Majority of the female applicants had Arts background whereas male applicants had Science background.
• There is almost a difference of 1% between pass and fail between the two genders. Female candidates for more unsuccessful vis-à-vis to the male candidates.
• Mathematics was more popular amongst male candidates than it was in female candidates.
Analysis of Column program
Program wise analysis shows that
• BBA and BSACF maintain some form of stability in the number of applicants received each year. BSCS witnessed a dramatic drop in 2017 followed by and exponential rise.
• The percentage of male applicants in all programs have been greater than the percentage of female applicants. However, this is not the case with SSLA as the percentage of female applicants is more than two times that of male candidates.
• BSCS is most popular in Sindh, while Punjab, Islamabad & Balochistan prefer BSECO as their most sorted degree. AJK, Foreign & Gilgit are more tilted towards the BSEM program
• In all degree programs except BSCS, applicants from Karachi were above average as compared to the others across all degree programs. SSLA score the lowest in Peshawar & Quetta as compared to other cities. • A Level & Intermediate was the most popular form of higher education amongst applicants. Most A Level applicants preferred BSEC whereas the Intermediate lot preferred BSCS. SSLA was most popular in the Aga Khan Board and those from the Other boards tilted towards BSEM
• About 69% of the Arts students preferred BSECO whereas 92% of the Science students chose BSCS
• the percentage of unsuccessful applicants across all programs was greater than the percentage of successful ones.
• 99% of those who studied math in their higher education pursued this as their degree by opting for BSEM. Those with non-mathematical background considered BSECO as their preferred degree.
Analysis of Column Province • the year 2019 witnessed the greatest number of applicants from AJK and Gilgit. In addition to this, in the year 2014 Gilgit had one of the lowest number of applicants. • AJK displayed the highest percentage of male applicants and lowest percentage of female applicants. The highest number of female applicants were from Sindh. • The highest percentage of applicants across all program was for BBA applicants from Baluchistan followed by applications for BBA from Sindh. • The city Hyderabad has the lowest turnout from Sindh. • A-Level was the least preferred choice for higher education in Gilgit. On the other hand, intermediate had the highest number of applicants across board from Baluchistan. • Foreign was the only state where Arts was selected, all other provinces clearly focused on Science.
• across all provinces the rate of unsuccessful candidates was far greater than the rate of successful candidates. • more than half of the applicants across board studied math in their higher education.
Analysis of Column city
• The percentage of applicants across cities have seen a gradual rise except for a slight dip in 2017 in the number of applicants from all cities.
• The percentage of male applicants was the highest from Peshawar whereas most female applicants were from Sindh.
• The highest percentage of applicants was witnessed for BBA degree from Quetta whereas the lowest popularity was of BSECO in Hyderabad.
• Aga Khan Board was the least popular in Lahore whereas, Intermediate had the highest percentage in Quetta.
• Arts was most popular in Sindh’s capital hub Karachi whereas, Science was most popular amongst Others and Quetta.
• As blatantly evident the percentage of successful candidates were far below the percentage of unsuccessful candidates.
• More than 50% of the applicants across all cities studied mathematics beforehand.
Analysis of Column cert_degree2
• The Aga Khan Board saw the highest percentage in 2017. Ina addition to this, the number of applicants in Other Boards have gradually increased.
• Percentage of Male candidates was highest from Intermediate, whereas most female candidates were from Aga Khan Board. • BBA was clearly the top scorer amongst claimants from all the different types of higher education certification. • A-Level & Intermediate board had the highest percentage amongst candidates from Sindh. Whereas, A-level in particular was the least preferred mode of higher education in AJK & Gilgit. • The Aga Khan Board scored the highest percentage in Karachi 3 and lowest in Lahore. Similarly, A-Level had the highest percentage in Karachi 3 and lowest in Quetta. • Arts had the highest percentage amongst candidates who did A Level whereas Science was most popular amongst those who did the Intermediate Board. • The percentage of unsuccessful candidates is greater than successful candidates across all higher education degree certification. • 80% of Aga Khan Board candidates studied mathematics Analysis of Column discipline2
• Arts was a popular discipline amongst female candidates whereas, the male candidates preferred Science.
• BBA was the most sought after degree for candidates having both Arts & Science background.
• Sindh had the highest percentage of candidates from both Arts & Science background. Arts scored the lowest in AJK & Gilgit.
• Karachi 3 had the highest percentage of both Arts & Science candidates. Lowest percentages were evident for Arts candidates in Peshawar & Quetta.
• Candidates perusing Arts & Science both were most likely to fall in for A-level higher education degree certification.
• Lie evident across other variables, in this case too, the percentage of unsuccessful candidates was far greater than those that were successful.
• those who selected the Arts discipline did not study math whereas those who took Science studied math earlier
Analysis of Column test_successful
• the percentage of unsuccessful candidates was the highest in 2019 and the percentage of successful candidates was the highest in 2018.
• 65% of the Male candidates were successful whereas only 34% of the female candidates were successful.
• BSCS & BSEM are the only variances whereby the percentage of successful candidate is greater than the percentage of unsuccessful candidates
• The highest percentage highlighted was the success percentage in Sindh (88%) whereas the lowest percentage of success rate was from AJK
• The highest number of unsuccessful candidates were from Karachi 3 whereas Peshawar had the lowest rate of unsuccessful candidates.
• the highest percentage of successful candidates were those with A Level background. On the contrary the highest percentage of unsuccessful candidates were from the Intermediate Board.
• those who had Science background managed to have a greater success rate than those with Arts background.
• Those who studied math had a greater success rate than those who didn’t.
Analysis of Column studied Mathematics • Over the years those who studied math is decreasing whereas percentage of those candidates who did not have math is increasing. • 69% of the male candidate had taken math in their higher education whereas only 30% of the female candidates had studied math. • Most candidates with or without mathematical background applied for the BA degree program. • The highest percentage of those who did not study math was from Sindh, whereas, the lowest percentage of those who did not study math was from AJK • the highest percentage of those who did not study math were from Karachi. On the other hand, the lowest percentage of those of did not study math were from Quetta. • Most A-Level candidates did not study math, however, most of Intermediate candidates did study math earlier on. • Those from Arts background did not study math but those with Science background had mathematical eloquence. • The percentage of unsuccessful applicants was greater than successful candidates irrespective of non-math versus mathematical background.
In [ ]: